id large_string | type large_string | ref_images list | ori_question large_string | options list | frames_list list | video_list list | system_prompt large_string | task_prompt large_string | user_prompt large_string | format_prompt large_string | ground_truth large_string |
|---|---|---|---|---|---|---|---|---|---|---|---|
question_0000 | Planning | [] | This video is non-sequentially filmed in a building, and it is known that there are only two toilets and one staircase inside this building. Currently, I am facing a large gray refrigerator. I want to go to the toilet and need to plan a route to reach the toilet as quickly as possible. What are the correct steps to tak... | [
"Turn right, go straight into the new room, then turn right and walk out of the room, open the door opposite the staircase and go in.",
"Turn right, go straight into the new room, then turn right and walk out of the room to reach the staircase and go upstairs.",
"Turn left, go straight into the new room, then t... | [
[
"question_0000/realestate10k_0000_0/00:37.50_frame_1.50_25.jpg",
"question_0000/realestate10k_0000_0/00:39.00_frame_1.50_26.jpg",
"question_0000/realestate10k_0000_0/00:40.50_frame_1.50_27.jpg",
"question_0000/realestate10k_0000_0/00:42.00_frame_1.50_28.jpg",
"question_0000/realestate10k_0000_... | [
{
"base_fps": 0.6666666666666666,
"end": 213,
"path": "question_0000/realestate10k_0000_0.mp4",
"start": 37.5
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video is non-sequentially filmed in a building, and it is known that there are only two toilets and one staircase inside this building. Currently, I am facing a large gray refrigerator. I want to go... | Please select the correct answer from the following options:
A. Turn right, go straight into the new room, then turn right and walk out of the room, open the door opposite the staircase and go in.;
B. Turn right, go straight into the new room, then turn right and walk out of the room to reach the staircase and go ups... | C |
question_0001 | Planning | [] | This video focuses on filming a cottage in the suburbs. After I go up the steps and enter the house through the door in front of me, if I want to cook something to eat, which way should I go? | [
"Open the door and enter the first room. Leave the room from the right, and turn left to reach the destination.",
"Open the door and enter the first room. Leave the room from the left, and turn right to reach the destination.",
"Open the door and enter the first room. Leave the room from the left, and enter the... | [
[
"question_0001/realestate10k_0006_0/00:06.00_frame_1.50_4.jpg",
"question_0001/realestate10k_0006_0/00:07.50_frame_1.50_5.jpg",
"question_0001/realestate10k_0006_0/00:09.00_frame_1.50_6.jpg",
"question_0001/realestate10k_0006_0/00:10.50_frame_1.50_7.jpg",
"question_0001/realestate10k_0006_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 180,
"path": "question_0001/realestate10k_0006_0.mp4",
"start": 6
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video focuses on filming a cottage in the suburbs. After I go up the steps and enter the house through the door in front of me, if I want to cook something to eat, which way should I go? | Please select the correct answer from the following options:
A. Open the door and enter the first room. Leave the room from the right, and turn left to reach the destination.;
B. Open the door and enter the first room. Leave the room from the left, and turn right to reach the destination.;
C. Open the door and enter ... | B |
question_0002 | Planning | [] | The video captures the entire process of my ride from the 0-second mark (the last frame is taken now). I suddenly realize that I left something at the starting point. How should I proceed to retrieve it? | [
"Turn around and keep riding forward until you pass 6 streetlights on the left, then turn left",
"Turn around and keep riding forward until you pass 2 streetlights on the left, then turn right",
"Turn around and keep riding forward until you pass 4 streetlights on the left, then turn right",
"Turn around and ... | [
[
"question_0002/ego4d_0094_0/00:00.00_frame_0.50_0.jpg",
"question_0002/ego4d_0094_0/00:00.50_frame_0.50_1.jpg",
"question_0002/ego4d_0094_0/00:01.00_frame_0.50_2.jpg",
"question_0002/ego4d_0094_0/00:01.50_frame_0.50_3.jpg",
"question_0002/ego4d_0094_0/00:02.00_frame_0.50_4.jpg",
"question_... | [
{
"base_fps": 2,
"end": 80.5,
"path": "question_0002/ego4d_0094_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video captures the entire process of my ride from the 0-second mark (the last frame is taken now). I suddenly realize that I left something at the starting point. How should I proceed to retrieve it? | Please select the correct answer from the following options:
A. Turn around and keep riding forward until you pass 6 streetlights on the left, then turn left;
B. Turn around and keep riding forward until you pass 2 streetlights on the left, then turn right;
C. Turn around and keep riding forward until you pass 4 stre... | C |
question_0003 | Planning | [] | Alex is the driver of this car, and this is the video footage captured by a camera on his car (the last frame represents the current moment). If Alex wants to return to the parking lot where he started, how should he drive now? | [
"U-turn and go straight, then turn left at the second intersection to enter the parking lot",
"U-turn and go straight, then turn left at the fourth intersection to enter the parking lot",
"U-turn and go straight, then turn right at the fourth intersection to enter the parking lot",
"U-turn and go straight, th... | [
[
"question_0003/nuscenes_0000_4/00:12.75_frame_0.25_51.jpg",
"question_0003/nuscenes_0000_4/00:13.00_frame_0.25_52.jpg",
"question_0003/nuscenes_0000_4/00:13.25_frame_0.25_53.jpg",
"question_0003/nuscenes_0000_4/00:13.50_frame_0.25_54.jpg",
"question_0003/nuscenes_0000_4/00:13.75_frame_0.25_55.... | [
{
"base_fps": 4.0001600064002565,
"end": 41.49833999999999,
"path": "question_0003/nuscenes_0000_4.mp4",
"start": 12.749489999999998
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Alex is the driver of this car, and this is the video footage captured by a camera on his car (the last frame represents the current moment). If Alex wants to return to the parking lot where he started, ... | Please select the correct answer from the following options:
A. U-turn and go straight, then turn left at the second intersection to enter the parking lot;
B. U-turn and go straight, then turn left at the fourth intersection to enter the parking lot;
C. U-turn and go straight, then turn right at the fourth intersecti... | D |
question_0004 | Planning | [
"question_0004/image_content_1.png"
] | The video is shot from a bird's-eye view of a building. I will provide you with an image showing my position and orientation on the ground (the red dot represents my position and the arrow shows my orientation)<image>Please answer: If I want to view the statue from the front, how many paths are there to reach the targe... | [
"3 paths; the shortest path is to enter the building from the rear-right passage and walk straight to the target point",
"2 paths; the shortest path is to go up the steps, turn right, and walk around the side of the statue to reach the target point",
"3 paths; the shortest path is to go up the steps, turn right... | [
[
"question_0004/dl3dv_0015_0/00:00.00_frame_1.00_0.jpg",
"question_0004/dl3dv_0015_0/00:01.00_frame_1.00_1.jpg",
"question_0004/dl3dv_0015_0/00:02.00_frame_1.00_2.jpg",
"question_0004/dl3dv_0015_0/00:03.00_frame_1.00_3.jpg",
"question_0004/dl3dv_0015_0/00:04.00_frame_1.00_4.jpg",
"question_... | [
{
"base_fps": 1,
"end": 30,
"path": "question_0004/dl3dv_0015_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video is shot from a bird's-eye view of a building. I will provide you with an image showing my position and orientation on the ground (the red dot represents my position and the arrow shows my orien... | Please select the correct answer from the following options:
A. 3 paths; the shortest path is to enter the building from the rear-right passage and walk straight to the target point;
B. 2 paths; the shortest path is to go up the steps, turn right, and walk around the side of the statue to reach the target point;
C. 3... | A |
question_0005 | Planning | [] | The video records my environment in the office, and the last frame represents my current position. Assuming it's raining heavily outside and I want to go out, which way should I go next? | [
"Turn right, walk straight to the doorknob position, pick up the thing in the corner, and open the door.",
"Turn right, walk straight to the end, turn left, walk straight to the end, pick up the thing on the ground, turn around, walk straight to the doorknob position and open the door.",
"Turn right, walk strai... | [
[
"question_0005/scannetpp_0104_0/00:00.00_frame_0.50_0.jpg",
"question_0005/scannetpp_0104_0/00:00.50_frame_0.50_1.jpg",
"question_0005/scannetpp_0104_0/00:01.00_frame_0.50_2.jpg",
"question_0005/scannetpp_0104_0/00:01.50_frame_0.50_3.jpg",
"question_0005/scannetpp_0104_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 110,
"path": "question_0005/scannetpp_0104_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video records my environment in the office, and the last frame represents my current position. Assuming it's raining heavily outside and I want to go out, which way should I go next? | Please select the correct answer from the following options:
A. Turn right, walk straight to the doorknob position, pick up the thing in the corner, and open the door.;
B. Turn right, walk straight to the end, turn left, walk straight to the end, pick up the thing on the ground, turn around, walk straight to the door... | B |
question_0006 | Planning | [] | Bob is the owner of this room. He is now sitting on the red sofa closest to the square table, and he wants to go to the green bed to sleep. How should he go? | [
"Turn left and walk a short distance to go upstairs, after going upstairs, enter the second room on the right",
"Turn left and walk a short distance to go upstairs, after going upstairs, enter the first room on the right",
"Walk forward for a distance, then turn left into the kitchen, continue forward out of th... | [
[
"question_0006/realestate10k_0044_0/00:07.50_frame_1.50_5.jpg",
"question_0006/realestate10k_0044_0/00:09.00_frame_1.50_6.jpg",
"question_0006/realestate10k_0044_0/00:10.50_frame_1.50_7.jpg",
"question_0006/realestate10k_0044_0/00:12.00_frame_1.50_8.jpg",
"question_0006/realestate10k_0044_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 82.5,
"path": "question_0006/realestate10k_0044_0.mp4",
"start": 7.5
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Bob is the owner of this room. He is now sitting on the red sofa closest to the square table, and he wants to go to the green bed to sleep. How should he go? | Please select the correct answer from the following options:
A. Turn left and walk a short distance to go upstairs, after going upstairs, enter the second room on the right;
B. Turn left and walk a short distance to go upstairs, after going upstairs, enter the first room on the right;
C. Walk forward for a distance, ... | A |
question_0007 | Planning | [
"question_0007/image_content_1.png"
] | This video shows the interior of a building (it is known that there is only one staircase in this building). Assuming I am sitting here at the dining table as shown in this image: <image> I am sitting in the corner. If I want to go upstairs, which of the following paths is correct and shortest? | [
"Exit the kitchen through the door currently directly in front of you, turn right after exiting, walk a short distance, then turn right around the corner to go straight upstairs",
"Exit the kitchen through the door currently on your left, turn left after exiting, walk a short distance to reach the carpeted area, ... | [
[
"question_0007/realestate10k_0044_0/00:00.00_frame_1.50_0.jpg",
"question_0007/realestate10k_0044_0/00:01.50_frame_1.50_1.jpg",
"question_0007/realestate10k_0044_0/00:03.00_frame_1.50_2.jpg",
"question_0007/realestate10k_0044_0/00:04.50_frame_1.50_3.jpg",
"question_0007/realestate10k_0044_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 145.5,
"path": "question_0007/realestate10k_0044_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video shows the interior of a building (it is known that there is only one staircase in this building). Assuming I am sitting here at the dining table as shown in this image: <image> I am sitting in... | Please select the correct answer from the following options:
A. Exit the kitchen through the door currently directly in front of you, turn right after exiting, walk a short distance, then turn right around the corner to go straight upstairs;
B. Exit the kitchen through the door currently on your left, turn left after... | C |
question_0008 | Planning | [] | This video records my walk around the gym, and the last frame represents my current position. If I want to leave to find my friend who is wearing green and running, how should I go? | [
"Rotate counterclockwise about 135 degrees, walk straight to the end of the wall, then turn left, open the door on the left and enter",
"Rotate counterclockwise about 135 degrees, walk straight to the end of the wall, then turn right, open the door on the right and enter",
"Rotate clockwise about 135 degrees, w... | [
[
"question_0008/dl3dv-10k_0012_0/00:00.00_frame_1.00_0.jpg",
"question_0008/dl3dv-10k_0012_0/00:01.00_frame_1.00_1.jpg",
"question_0008/dl3dv-10k_0012_0/00:02.00_frame_1.00_2.jpg",
"question_0008/dl3dv-10k_0012_0/00:03.00_frame_1.00_3.jpg",
"question_0008/dl3dv-10k_0012_0/00:04.00_frame_1.00_4.... | [
{
"base_fps": 1,
"end": 87,
"path": "question_0008/dl3dv-10k_0012_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video records my walk around the gym, and the last frame represents my current position. If I want to leave to find my friend who is wearing green and running, how should I go? | Please select the correct answer from the following options:
A. Rotate counterclockwise about 135 degrees, walk straight to the end of the wall, then turn left, open the door on the left and enter;
B. Rotate counterclockwise about 135 degrees, walk straight to the end of the wall, then turn right, open the door on th... | B |
question_0009 | Planning | [] | I am working in the room, and I suddenly want to closely admire the striking red painting. How should I proceed? | [
"Turn around, walk straight, and enter after passing the first door on your left",
"Turn around and walk straight to the end",
"Turn around, walk straight, and turn right at the first intersection",
"Go left and walk straight until you reach the sofa"
] | [
[
"question_0009/scannet_fix_0069_0/00:00.00_frame_1.00_0.jpg",
"question_0009/scannet_fix_0069_0/00:01.00_frame_1.00_1.jpg",
"question_0009/scannet_fix_0069_0/00:02.00_frame_1.00_2.jpg",
"question_0009/scannet_fix_0069_0/00:03.00_frame_1.00_3.jpg",
"question_0009/scannet_fix_0069_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 108,
"path": "question_0009/scannet_fix_0069_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am working in the room, and I suddenly want to closely admire the striking red painting. How should I proceed? | Please select the correct answer from the following options:
A. Turn around, walk straight, and enter after passing the first door on your left;
B. Turn around and walk straight to the end;
C. Turn around, walk straight, and turn right at the first intersection;
D. Go left and walk straight until you reach the sofa.
... | B |
question_0010 | Planning | [
"question_0010/image_content_2.png",
"question_0010/image_content_4.png"
] | I am currently working in the position shown in image one within the room displayed in the video; I want to get some fresh air on the balcony shown in image two. Please help me plan a route. Image one: <image> Image two: <image> | [
"Enter the dining area from the corridor on the left, turn left after entering, go straight to the end, and enter the door on the left.",
"Enter the living room from the corridor on the right, turn right after entering, go straight to the end, then turn left, and keep going forward to enter another door.",
"Ent... | [
[
"question_0010/realestate10k_0014_0/00:00.00_frame_1.50_0.jpg",
"question_0010/realestate10k_0014_0/00:01.50_frame_1.50_1.jpg",
"question_0010/realestate10k_0014_0/00:03.00_frame_1.50_2.jpg",
"question_0010/realestate10k_0014_0/00:04.50_frame_1.50_3.jpg",
"question_0010/realestate10k_0014_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 303,
"path": "question_0010/realestate10k_0014_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am currently working in the position shown in image one within the room displayed in the video; I want to get some fresh air on the balcony shown in image two. Please help me plan a route. Image one: <... | Please select the correct answer from the following options:
A. Enter the dining area from the corridor on the left, turn left after entering, go straight to the end, and enter the door on the left.;
B. Enter the living room from the corridor on the right, turn right after entering, go straight to the end, then turn ... | C |
question_0011 | Planning | [
"question_0011/image_content_1.png",
"question_0011/image_content_4.png",
"question_0011/image_content_6.png"
] | This video is a disjointed recording of a building. I now want to go to the kitchen to cook, starting from the location shown in the image: <image> Please determine how I should proceed at the following two intersections on the way to the kitchen. Intersection one: <image> Intersection two: <image> | [
"At intersection one, turn around and go straight for a distance then turn right and go downstairs; at intersection two, turn around and go straight to exit the room",
"At intersection one, turn around and go straight for a distance then turn left and go downstairs; at intersection two, turn around and go straigh... | [
[
"question_0011/realestate10k_0000_0/00:00.00_frame_1.50_0.jpg",
"question_0011/realestate10k_0000_0/00:01.50_frame_1.50_1.jpg",
"question_0011/realestate10k_0000_0/00:03.00_frame_1.50_2.jpg",
"question_0011/realestate10k_0000_0/00:04.50_frame_1.50_3.jpg",
"question_0011/realestate10k_0000_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 228,
"path": "question_0011/realestate10k_0000_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video is a disjointed recording of a building. I now want to go to the kitchen to cook, starting from the location shown in the image: <image> Please determine how I should proceed at the following ... | Please select the correct answer from the following options:
A. At intersection one, turn around and go straight for a distance then turn right and go downstairs; at intersection two, turn around and go straight to exit the room;
B. At intersection one, turn around and go straight for a distance then turn left and go... | C |
question_0012 | Planning | [] | The video shows the layout of a room, and I am currently sitting on a blue-gray chair that appears at the 53rd second. If I want to leave this room as quickly as possible, please help me plan the shortest route. | [
"Turn directly to the right, walk straight to the door, and open it to go out",
"Turn directly to the left, walk straight to the door, and open it to go out",
"Turn directly backward, walk straight to the door, and open it to go out",
"Go directly forward, walk around the table and continue straight to the do... | [
[
"question_0012/scannet_fix_0037_0/00:00.00_frame_1.00_0.jpg",
"question_0012/scannet_fix_0037_0/00:01.00_frame_1.00_1.jpg",
"question_0012/scannet_fix_0037_0/00:02.00_frame_1.00_2.jpg",
"question_0012/scannet_fix_0037_0/00:03.00_frame_1.00_3.jpg",
"question_0012/scannet_fix_0037_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 83,
"path": "question_0012/scannet_fix_0037_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video shows the layout of a room, and I am currently sitting on a blue-gray chair that appears at the 53rd second. If I want to leave this room as quickly as possible, please help me plan the shortes... | Please select the correct answer from the following options:
A. Turn directly to the right, walk straight to the door, and open it to go out;
B. Turn directly to the left, walk straight to the door, and open it to go out;
C. Turn directly backward, walk straight to the door, and open it to go out;
D. Go directly forw... | C |
question_0013 | Planning | [
"question_0013/image_content_1.png",
"question_0013/image_content_3.png"
] | This video records the interior layout of a building, and this image represents my current position and orientation: <image> I want to reach this dining table (as shown in the image below), what is the shortest path (select the correct option)? <image> | [
"Walk straight for a distance and turn left to go down the stairs, after going down the stairs, turn around and enter the right corridor, walk straight for a distance",
"Walk straight for a distance and go straight down the stairs, after going down the stairs, turn around and enter the right corridor, walk straig... | [
[
"question_0013/realestate10k_0000_0/00:00.00_frame_1.50_0.jpg",
"question_0013/realestate10k_0000_0/00:01.50_frame_1.50_1.jpg",
"question_0013/realestate10k_0000_0/00:03.00_frame_1.50_2.jpg",
"question_0013/realestate10k_0000_0/00:04.50_frame_1.50_3.jpg",
"question_0013/realestate10k_0000_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 180,
"path": "question_0013/realestate10k_0000_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video records the interior layout of a building, and this image represents my current position and orientation: <image> I want to reach this dining table (as shown in the image below), what is the s... | Please select the correct answer from the following options:
A. Walk straight for a distance and turn left to go down the stairs, after going down the stairs, turn around and enter the right corridor, walk straight for a distance;
B. Walk straight for a distance and go straight down the stairs, after going down the s... | C |
question_0014 | Planning | [] | The length and width of this room are x and y respectively (x is greater than y). I currently have a book in hand and want to sit on a swivel chair in a sunlit spot to read comfortably. I entered this room from the door with a purple mark. What should I do next? | [
"Turn left along the pool after entering, walk straight about 0.5x, then turn left to reach the chair.",
"Turn left along the pool after entering, walk straight until you encounter an obstacle, turn left to reach the chair.",
"Walk forward about 0.5y, then turn left, then walk straight about 0.5x, turn right an... | [
[
"question_0014/scannetpp_0009_0/00:50.00_frame_0.50_100.jpg",
"question_0014/scannetpp_0009_0/00:50.50_frame_0.50_101.jpg",
"question_0014/scannetpp_0009_0/00:51.00_frame_0.50_102.jpg",
"question_0014/scannetpp_0009_0/00:51.50_frame_0.50_103.jpg",
"question_0014/scannetpp_0009_0/00:52.00_frame... | [
{
"base_fps": 2,
"end": 137.5,
"path": "question_0014/scannetpp_0009_0.mp4",
"start": 50
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The length and width of this room are x and y respectively (x is greater than y). I currently have a book in hand and want to sit on a swivel chair in a sunlit spot to read comfortably. I entered this ro... | Please select the correct answer from the following options:
A. Turn left along the pool after entering, walk straight about 0.5x, then turn left to reach the chair.;
B. Turn left along the pool after entering, walk straight until you encounter an obstacle, turn left to reach the chair.;
C. Walk forward about 0.5y, t... | D |
question_0015 | Planning | [] | Suppose I am currently sitting on that white chair closest to the cabinet, and I want to operate the robotic arm to conduct an experiment. How should I proceed? | [
"Get up, turn left, walk forward a certain distance, then turn left again, and walk a certain distance to reach the target point.",
"Get up, turn left, walk forward a certain distance, then turn right, and walk a certain distance to reach the target point.",
"Get up, walk backward a certain distance, then turn ... | [
[
"question_0015/scannetpp_0002_0/00:45.00_frame_0.50_90.jpg",
"question_0015/scannetpp_0002_0/00:45.50_frame_0.50_91.jpg",
"question_0015/scannetpp_0002_0/00:46.00_frame_0.50_92.jpg",
"question_0015/scannetpp_0002_0/00:46.50_frame_0.50_93.jpg",
"question_0015/scannetpp_0002_0/00:47.00_frame_0.5... | [
{
"base_fps": 2,
"end": 140,
"path": "question_0015/scannetpp_0002_0.mp4",
"start": 45
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Suppose I am currently sitting on that white chair closest to the cabinet, and I want to operate the robotic arm to conduct an experiment. How should I proceed? | Please select the correct answer from the following options:
A. Get up, turn left, walk forward a certain distance, then turn left again, and walk a certain distance to reach the target point.;
B. Get up, turn left, walk forward a certain distance, then turn right, and walk a certain distance to reach the target poin... | A |
question_0016 | Planning | [
"question_0016/image_content_1.png"
] | I am attending a class in the classroom shown in the video. The green dot and arrow on the floor represent my current position and orientation: <image>. I now need to use the remote control to turn on the projector. What is the shortest path I should take? (Please select the correct and shortest path.). Hint: The long ... | [
"Turn around, walk straight about 4.8m, the remote control is on your left hand side",
"Turn right around the table to the other side, walk straight about 4.8m, the remote control is on your right hand side",
"Turn around, walk straight about 1.6m, the remote control is on your left hand side",
"Turn right ar... | [
[
"question_0016/scannet_fix_0053_0/00:00.00_frame_1.00_0.jpg",
"question_0016/scannet_fix_0053_0/00:01.00_frame_1.00_1.jpg",
"question_0016/scannet_fix_0053_0/00:02.00_frame_1.00_2.jpg",
"question_0016/scannet_fix_0053_0/00:03.00_frame_1.00_3.jpg",
"question_0016/scannet_fix_0053_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 54,
"path": "question_0016/scannet_fix_0053_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am attending a class in the classroom shown in the video. The green dot and arrow on the floor represent my current position and orientation: <image>. I now need to use the remote control to turn on th... | Please select the correct answer from the following options:
A. Turn around, walk straight about 4.8m, the remote control is on your left hand side;
B. Turn right around the table to the other side, walk straight about 4.8m, the remote control is on your right hand side;
C. Turn around, walk straight about 1.6m, the ... | A |
question_0017 | Planning | [] | Observe the video filmed in the library. It seems I left a black jacket here. In front of me, there are several red books stacked side by side, and behind me, there are some computers. How should I proceed to retrieve my jacket? | [
"Turn right, walk straight to the end of this bookshelf, then turn right and you'll see it",
"Turn left, walk straight to the end of this bookshelf, then turn left and you'll see it",
"Turn left, walk straight to the end of this bookshelf, then turn right and you'll see it",
"The jacket does not appear in the... | [
[
"question_0017/scannet_fix_0052_0/00:00.00_frame_1.00_0.jpg",
"question_0017/scannet_fix_0052_0/00:01.00_frame_1.00_1.jpg",
"question_0017/scannet_fix_0052_0/00:02.00_frame_1.00_2.jpg",
"question_0017/scannet_fix_0052_0/00:03.00_frame_1.00_3.jpg",
"question_0017/scannet_fix_0052_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 38,
"path": "question_0017/scannet_fix_0052_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Observe the video filmed in the library. It seems I left a black jacket here. In front of me, there are several red books stacked side by side, and behind me, there are some computers. How should I proce... | Please select the correct answer from the following options:
A. Turn right, walk straight to the end of this bookshelf, then turn right and you'll see it;
B. Turn left, walk straight to the end of this bookshelf, then turn left and you'll see it;
C. Turn left, walk straight to the end of this bookshelf, then turn rig... | A |
question_0018 | Planning | [
"question_0018/image_content_1.png"
] | This video captures the interior of a building. I am currently inside this building, and the following image represents the view in front of me. I want to go to the laundry room to pick up my clothes. What should I do next? <image> | [
"Turn around and exit the kitchen, walk left for a distance",
"Turn around and exit the kitchen, walk right for a distance",
"Turn right from the front-right along the white kitchen counter, exit the kitchen, and then walk right for a distance",
"Walk straight for a distance, exit the kitchen through the wood... | [
[
"question_0018/realestate10k_0003_0/00:06.00_frame_1.50_4.jpg",
"question_0018/realestate10k_0003_0/00:07.50_frame_1.50_5.jpg",
"question_0018/realestate10k_0003_0/00:09.00_frame_1.50_6.jpg",
"question_0018/realestate10k_0003_0/00:10.50_frame_1.50_7.jpg",
"question_0018/realestate10k_0003_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 96,
"path": "question_0018/realestate10k_0003_0.mp4",
"start": 6
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video captures the interior of a building. I am currently inside this building, and the following image represents the view in front of me. I want to go to the laundry room to pick up my clothes. Wh... | Please select the correct answer from the following options:
A. Turn around and exit the kitchen, walk left for a distance;
B. Turn around and exit the kitchen, walk right for a distance;
C. Turn right from the front-right along the white kitchen counter, exit the kitchen, and then walk right for a distance;
D. Walk ... | A |
question_0019 | Planning | [] | This video captures the interior layout of a building. I am currently lying on a large double bed, resting. To go outside for some fresh air, how should I proceed? | [
"Leave the bedroom, go through one wooden door, then walk straight to the right into the kitchen, exit the kitchen to reach the location of the main door.",
"Leave the bedroom, go through two wooden doors, then move to the front-left to enter the living room, exit the living room to reach the location of the main... | [
[
"question_0019/realestate10k_0003_0/00:06.00_frame_1.50_4.jpg",
"question_0019/realestate10k_0003_0/00:07.50_frame_1.50_5.jpg",
"question_0019/realestate10k_0003_0/00:09.00_frame_1.50_6.jpg",
"question_0019/realestate10k_0003_0/00:10.50_frame_1.50_7.jpg",
"question_0019/realestate10k_0003_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 135,
"path": "question_0019/realestate10k_0003_0.mp4",
"start": 6
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video captures the interior layout of a building. I am currently lying on a large double bed, resting. To go outside for some fresh air, how should I proceed? | Please select the correct answer from the following options:
A. Leave the bedroom, go through one wooden door, then walk straight to the right into the kitchen, exit the kitchen to reach the location of the main door.;
B. Leave the bedroom, go through two wooden doors, then move to the front-left to enter the living ... | D |
question_0020 | Prediction | [
"question_0020/image_content_0.png",
"question_0020/image_content_2.png"
] | <image>This video records my driving so far. Assuming the average speed of the woman shown remains constant and her walking direction doesn't change, where is she approximately located at present?<image> | [
"red dot position",
"green dot position",
"pink dot position",
"yellow dot position"
] | [
[
"question_0020/waymo_0002_2/00:08.00_frame_0.20_40.jpg",
"question_0020/waymo_0002_2/00:08.20_frame_0.20_41.jpg",
"question_0020/waymo_0002_2/00:08.40_frame_0.20_42.jpg",
"question_0020/waymo_0002_2/00:08.60_frame_0.20_43.jpg",
"question_0020/waymo_0002_2/00:08.80_frame_0.20_44.jpg",
"ques... | [
{
"base_fps": 5,
"end": 18.8,
"path": "question_0020/waymo_0002_2.mp4",
"start": 8
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image>This video records my driving so far. Assuming the average speed of the woman shown remains constant and her walking direction doesn't change, where is she approximately located at present?<image> | Please select the correct answer from the following options:
A. red dot position;
B. green dot position;
C. pink dot position;
D. yellow dot position.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Onl... | B |
question_0021 | Prediction | [] | This video captures the remarkable performances of three athletes (the last frame represents the current moment). Please determine if the two athletes on the left and right will jump forward or backward in the next second (facing the camera is considered forward). | [
"Both athletes jump backward",
"The athlete on the left jumps backward, the athlete on the right jumps forward",
"The athlete on the left jumps forward, the athlete on the right jumps backward",
"Both athletes jump forward"
] | [
[
"question_0021/multisports_0082_0/00:00.00_frame_0.50_0.jpg",
"question_0021/multisports_0082_0/00:00.50_frame_0.50_1.jpg",
"question_0021/multisports_0082_0/00:01.00_frame_0.50_2.jpg",
"question_0021/multisports_0082_0/00:01.50_frame_0.50_3.jpg",
"question_0021/multisports_0082_0/00:02.00_fra... | [
{
"base_fps": 2,
"end": 20.5,
"path": "question_0021/multisports_0082_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video captures the remarkable performances of three athletes (the last frame represents the current moment). Please determine if the two athletes on the left and right will jump forward or backward ... | Please select the correct answer from the following options:
A. Both athletes jump backward;
B. The athlete on the left jumps backward, the athlete on the right jumps forward;
C. The athlete on the left jumps forward, the athlete on the right jumps backward;
D. Both athletes jump forward.
. You need to choose one let... | A |
question_0022 | Prediction | [] | This video shows the actions of three gymnasts. Please predict, in the next second of the video, in which direction will these three gymnasts do a somersault (facing the camera as the front)? | [
"From the left gymnast to the right gymnast: forward, backward, backward",
"From the left gymnast to the right gymnast: forward, backward, forward",
"From the left gymnast to the right gymnast: backward, backward, forward",
"From the left gymnast to the right gymnast: forward, forward, forward"
] | [
[
"question_0022/multisports_0039_0/00:00.00_frame_0.50_0.jpg",
"question_0022/multisports_0039_0/00:00.50_frame_0.50_1.jpg",
"question_0022/multisports_0039_0/00:01.00_frame_0.50_2.jpg",
"question_0022/multisports_0039_0/00:01.50_frame_0.50_3.jpg",
"question_0022/multisports_0039_0/00:02.00_fra... | [
{
"base_fps": 2,
"end": 20.5,
"path": "question_0022/multisports_0039_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video shows the actions of three gymnasts. Please predict, in the next second of the video, in which direction will these three gymnasts do a somersault (facing the camera as the front)? | Please select the correct answer from the following options:
A. From the left gymnast to the right gymnast: forward, backward, backward;
B. From the left gymnast to the right gymnast: forward, backward, forward;
C. From the left gymnast to the right gymnast: backward, backward, forward;
D. From the left gymnast to th... | B |
question_0023 | Prediction | [] | This video records the trajectory of the white car up to now. Assuming the white car's future direction remains unchanged and its speed equals the average speed during this period, please predict approximately how many seconds it will take for the car to reach the location of the green strip sign? | [
"11.5s",
"13.5s",
"6.5s",
"9.5s"
] | [
[
"question_0023/uav123_0009_0/00:00.00_frame_0.17_0.jpg",
"question_0023/uav123_0009_0/00:00.17_frame_0.17_1.jpg",
"question_0023/uav123_0009_0/00:00.33_frame_0.17_2.jpg",
"question_0023/uav123_0009_0/00:00.50_frame_0.17_3.jpg",
"question_0023/uav123_0009_0/00:00.67_frame_0.17_4.jpg",
"ques... | [
{
"base_fps": 6.0006000600060005,
"end": 12.6654,
"path": "question_0023/uav123_0009_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video records the trajectory of the white car up to now. Assuming the white car's future direction remains unchanged and its speed equals the average speed during this period, please predict approxi... | Please select the correct answer from the following options:
A. 11.5s;
B. 13.5s;
C. 6.5s;
D. 9.5s.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON for... | D |
question_0024 | Prediction | [] | This is a video shot inside a room. Assuming you start from the position and orientation shown at 1 minute 7 seconds and rotate counterclockwise in place, what is the order in which these objects appear in your view: a vase with pink flowers, a window, and a television? | [
"a window, a vase with pink flowers, a television",
"a window, a television, a vase with pink flowers",
"a television, a vase with pink flowers, a window",
"a vase with pink flowers, a window, a television"
] | [
[
"question_0024/arkitscenes_fix_0039_0/00:00.00_frame_1.00_0.jpg",
"question_0024/arkitscenes_fix_0039_0/00:01.00_frame_1.00_1.jpg",
"question_0024/arkitscenes_fix_0039_0/00:02.00_frame_1.00_2.jpg",
"question_0024/arkitscenes_fix_0039_0/00:03.00_frame_1.00_3.jpg",
"question_0024/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 85,
"path": "question_0024/arkitscenes_fix_0039_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This is a video shot inside a room. Assuming you start from the position and orientation shown at 1 minute 7 seconds and rotate counterclockwise in place, what is the order in which these objects appear ... | Please select the correct answer from the following options:
A. a window, a vase with pink flowers, a television;
B. a window, a television, a vase with pink flowers;
C. a television, a vase with pink flowers, a window;
D. a vase with pink flowers, a window, a television.
. You need to choose one letter from A,B,C,D ... | B |
question_0025 | Prediction | [] | You have filmed these videos continuously in a room (the last frame is your current shot). Based on your current orientation and position, after performing the following operations, which object from the options will be closest to you: turn left 90 degrees, walk straight to the end, then turn right 90 degrees, walk str... | [
"rectangular bed board",
"white cabinet",
"computer",
"black speaker"
] | [
[
"question_0025/scannet_fix_0011_0/00:00.00_frame_1.00_0.jpg",
"question_0025/scannet_fix_0011_0/00:01.00_frame_1.00_1.jpg",
"question_0025/scannet_fix_0011_0/00:02.00_frame_1.00_2.jpg",
"question_0025/scannet_fix_0011_0/00:03.00_frame_1.00_3.jpg",
"question_0025/scannet_fix_0011_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 41,
"path": "question_0025/scannet_fix_0011_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: You have filmed these videos continuously in a room (the last frame is your current shot). Based on your current orientation and position, after performing the following operations, which object from the... | Please select the correct answer from the following options:
A. rectangular bed board;
B. white cabinet;
C. computer;
D. black speaker.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your ... | D |
question_0026 | Prediction | [
"question_0026/image_content_0.png"
] | <image> Based on video information, assuming you start at the position represented in the image, after each turn you will immediately walk straight until encountering an obstacle. After executing the following actions: turn left 90 degrees, turn left 90 degrees, turn left 90 degrees, and turn right 90 degrees, in which... | [
"front-right",
"rear-left",
"directly left",
"directly right"
] | [
[
"question_0026/dl3dv-10k_0094_0/00:00.00_frame_1.00_0.jpg",
"question_0026/dl3dv-10k_0094_0/00:01.00_frame_1.00_1.jpg",
"question_0026/dl3dv-10k_0094_0/00:02.00_frame_1.00_2.jpg",
"question_0026/dl3dv-10k_0094_0/00:03.00_frame_1.00_3.jpg",
"question_0026/dl3dv-10k_0094_0/00:04.00_frame_1.00_4.... | [
{
"base_fps": 1,
"end": 90,
"path": "question_0026/dl3dv-10k_0094_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image> Based on video information, assuming you start at the position represented in the image, after each turn you will immediately walk straight until encountering an obstacle. After executing the fol... | Please select the correct answer from the following options:
A. front-right;
B. rear-left;
C. directly left;
D. directly right.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(a... | D |
question_0027 | Prediction | [] | Assume you are sitting on the toilet. Starting from directly to your left and rotating clockwise, what do you see in order? (Select the correct sequence) | [
"Door, tissue dispenser, sink",
"Door, trash can, sink",
"Mirror, door, trash can",
"Trash can, mirror, handle"
] | [
[
"question_0027/scannet_fix_0014_0/00:00.00_frame_1.00_0.jpg",
"question_0027/scannet_fix_0014_0/00:01.00_frame_1.00_1.jpg",
"question_0027/scannet_fix_0014_0/00:02.00_frame_1.00_2.jpg",
"question_0027/scannet_fix_0014_0/00:03.00_frame_1.00_3.jpg",
"question_0027/scannet_fix_0014_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 25,
"path": "question_0027/scannet_fix_0014_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Assume you are sitting on the toilet. Starting from directly to your left and rotating clockwise, what do you see in order? (Select the correct sequence) | Please select the correct answer from the following options:
A. Door, tissue dispenser, sink;
B. Door, trash can, sink;
C. Mirror, door, trash can;
D. Trash can, mirror, handle.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arriv... | C |
question_0028 | Prediction | [
"question_0028/image_content_0.png"
] | <image> Assuming I start from the position of the red dot shown in the image and walk towards the direction of the door, how many objects will be on my right side: TV; microwave oven; wooden shelf; frying pan | [
"4",
"2",
"1",
"3"
] | [
[
"question_0028/arkitscenes_fix_0019_0/00:00.00_frame_1.00_0.jpg",
"question_0028/arkitscenes_fix_0019_0/00:01.00_frame_1.00_1.jpg",
"question_0028/arkitscenes_fix_0019_0/00:02.00_frame_1.00_2.jpg",
"question_0028/arkitscenes_fix_0019_0/00:03.00_frame_1.00_3.jpg",
"question_0028/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 49,
"path": "question_0028/arkitscenes_fix_0019_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image> Assuming I start from the position of the red dot shown in the image and walk towards the direction of the door, how many objects will be on my right side: TV; microwave oven; wooden shelf; fryin... | Please select the correct answer from the following options:
A. 4;
B. 2;
C. 1;
D. 3.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | C |
question_0029 | Prediction | [
"question_0029/image_content_0.png"
] | Assuming there is a person moving at a constant speed from the red point to the green point as shown in the image <image> at 2 seconds, and is expected to reach the green point at 14 seconds. Assuming the athlete does not notice this person and continues with their original motion, about when does the first collision h... | [
"7 seconds to 9 seconds",
"5 seconds to 7 seconds",
"No collision will occur",
"9 seconds to 11 seconds"
] | [
[
"question_0029/multisports_0012_0/00:00.00_frame_0.50_0.jpg",
"question_0029/multisports_0012_0/00:00.50_frame_0.50_1.jpg",
"question_0029/multisports_0012_0/00:01.00_frame_0.50_2.jpg",
"question_0029/multisports_0012_0/00:01.50_frame_0.50_3.jpg",
"question_0029/multisports_0012_0/00:02.00_fra... | [
{
"base_fps": 2,
"end": 40.5,
"path": "question_0029/multisports_0012_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Assuming there is a person moving at a constant speed from the red point to the green point as shown in the image <image> at 2 seconds, and is expected to reach the green point at 14 seconds. Assuming th... | Please select the correct answer from the following options:
A. 7 seconds to 9 seconds;
B. 5 seconds to 7 seconds;
C. No collision will occur;
D. 9 seconds to 11 seconds.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at t... | A |
question_0030 | (Spatial Construction) Instance/Scene Attribute | [] | How many basketball hoops are there in this sports field? | [
"4",
"5",
"3",
"2"
] | [
[
"question_0030/dl3dv_0006_0/00:00.00_frame_1.00_0.jpg",
"question_0030/dl3dv_0006_0/00:01.00_frame_1.00_1.jpg",
"question_0030/dl3dv_0006_0/00:02.00_frame_1.00_2.jpg",
"question_0030/dl3dv_0006_0/00:03.00_frame_1.00_3.jpg",
"question_0030/dl3dv_0006_0/00:04.00_frame_1.00_4.jpg",
"question_... | [
{
"base_fps": 1,
"end": 29,
"path": "question_0030/dl3dv_0006_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: How many basketball hoops are there in this sports field? | Please select the correct answer from the following options:
A. 4;
B. 5;
C. 3;
D. 2.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | A |
question_0031 | (Spatial Construction) Instance/Scene Attribute | [] | Assuming you are sitting on a brown bench and want to enter the interior of the teaching building, how many entrances are there (without climbing over railings/windows)? | [
"6",
"4",
"7",
"5"
] | [
[
"question_0031/dl3dv_0009_0/00:00.00_frame_1.00_0.jpg",
"question_0031/dl3dv_0009_0/00:01.00_frame_1.00_1.jpg",
"question_0031/dl3dv_0009_0/00:02.00_frame_1.00_2.jpg",
"question_0031/dl3dv_0009_0/00:03.00_frame_1.00_3.jpg",
"question_0031/dl3dv_0009_0/00:04.00_frame_1.00_4.jpg",
"question_... | [
{
"base_fps": 1,
"end": 24,
"path": "question_0031/dl3dv_0009_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Assuming you are sitting on a brown bench and want to enter the interior of the teaching building, how many entrances are there (without climbing over railings/windows)? | Please select the correct answer from the following options:
A. 6;
B. 4;
C. 7;
D. 5.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | D |
question_0032 | (Spatial Construction) Instance/Scene Attribute | [] | In this building, there is a chandelier with five lights, a blue pillow on the bed, and a rectangular painting with golden flowers. What is the order of their horizontal height from low to high? | [
"Chandelier with five lights, blue pillow on the bed, rectangular painting with golden flowers",
"Blue pillow on the bed, rectangular painting with golden flowers, chandelier with five lights",
"Rectangular painting with golden flowers, chandelier with five lights, blue pillow on the bed",
"Rectangular painti... | [
[
"question_0032/realestate10k_0004_0/00:15.00_frame_1.50_10.jpg",
"question_0032/realestate10k_0004_0/00:16.50_frame_1.50_11.jpg",
"question_0032/realestate10k_0004_0/00:18.00_frame_1.50_12.jpg",
"question_0032/realestate10k_0004_0/00:19.50_frame_1.50_13.jpg",
"question_0032/realestate10k_0004_... | [
{
"base_fps": 0.6666666666666666,
"end": 180,
"path": "question_0032/realestate10k_0004_0.mp4",
"start": 15
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: In this building, there is a chandelier with five lights, a blue pillow on the bed, and a rectangular painting with golden flowers. What is the order of their horizontal height from low to high? | Please select the correct answer from the following options:
A. Chandelier with five lights, blue pillow on the bed, rectangular painting with golden flowers;
B. Blue pillow on the bed, rectangular painting with golden flowers, chandelier with five lights;
C. Rectangular painting with golden flowers, chandelier with ... | A |
question_0033 | (Spatial Construction) Instance/Scene Attribute | [] | What is the distribution of restrooms inside the building shown in the video? | [
"Two on the first floor, one on the second floor",
"No restrooms on the first floor, two on the second floor",
"One on the first floor, two on the second floor",
"One on the first floor, one on the second floor"
] | [
[
"question_0033/realestate10k_0004_0/00:00.00_frame_1.50_0.jpg",
"question_0033/realestate10k_0004_0/00:01.50_frame_1.50_1.jpg",
"question_0033/realestate10k_0004_0/00:03.00_frame_1.50_2.jpg",
"question_0033/realestate10k_0004_0/00:04.50_frame_1.50_3.jpg",
"question_0033/realestate10k_0004_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 195,
"path": "question_0033/realestate10k_0004_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: What is the distribution of restrooms inside the building shown in the video? | Please select the correct answer from the following options:
A. Two on the first floor, one on the second floor;
B. No restrooms on the first floor, two on the second floor;
C. One on the first floor, two on the second floor;
D. One on the first floor, one on the second floor.
. You need to choose one letter from A,B... | C |
question_0034 | (Spatial Construction) Instance/Scene Attribute | [] | This video is filmed in a living room area, how many seating places are there in this living room area? | [
"5",
"3",
"4",
"6"
] | [
[
"question_0034/3rscan_0000_2/00:00.00_frame_1.00_0.jpg",
"question_0034/3rscan_0000_2/00:01.00_frame_1.00_1.jpg",
"question_0034/3rscan_0000_2/00:02.00_frame_1.00_2.jpg",
"question_0034/3rscan_0000_2/00:03.00_frame_1.00_3.jpg",
"question_0034/3rscan_0000_2/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1,
"end": 50,
"path": "question_0034/3rscan_0000_2.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video is filmed in a living room area, how many seating places are there in this living room area? | Please select the correct answer from the following options:
A. 5;
B. 3;
C. 4;
D. 6.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | D |
question_0035 | (Spatial Construction) Instance/Scene Attribute | [
"question_0035/image_content_0.png",
"question_0035/image_content_1.png"
] | <image><image> This video was shot inside a dining area. Did this apple and this dish appear in the room, and if so, which is higher in level height? | [
"Both appeared, and the dish is higher in level height",
"Both appeared, and the apple is higher in level height",
"At least one of the apple/dish did not appear in the room",
"Both appeared, and their level heights are about the same"
] | [
[
"question_0035/3rscan_0001_2/02:55.00_frame_1.00_175.jpg",
"question_0035/3rscan_0001_2/02:56.00_frame_1.00_176.jpg",
"question_0035/3rscan_0001_2/02:57.00_frame_1.00_177.jpg",
"question_0035/3rscan_0001_2/02:58.00_frame_1.00_178.jpg",
"question_0035/3rscan_0001_2/02:59.00_frame_1.00_179.jpg",... | [
{
"base_fps": 1,
"end": 275,
"path": "question_0035/3rscan_0001_2.mp4",
"start": 175
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image><image> This video was shot inside a dining area. Did this apple and this dish appear in the room, and if so, which is higher in level height? | Please select the correct answer from the following options:
A. Both appeared, and the dish is higher in level height;
B. Both appeared, and the apple is higher in level height;
C. At least one of the apple/dish did not appear in the room;
D. Both appeared, and their level heights are about the same.
. You need to ch... | A |
question_0036 | (Spatial Construction) Instance/Scene Attribute | [] | How many spaces with handles that can be pulled open are in the kitchen cooking area shown in the video? | [
"15",
"16",
"14",
"13"
] | [
[
"question_0036/3rscan_0001_0/00:00.00_frame_1.00_0.jpg",
"question_0036/3rscan_0001_0/00:01.00_frame_1.00_1.jpg",
"question_0036/3rscan_0001_0/00:02.00_frame_1.00_2.jpg",
"question_0036/3rscan_0001_0/00:03.00_frame_1.00_3.jpg",
"question_0036/3rscan_0001_0/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1,
"end": 70,
"path": "question_0036/3rscan_0001_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: How many spaces with handles that can be pulled open are in the kitchen cooking area shown in the video? | Please select the correct answer from the following options:
A. 15;
B. 16;
C. 14;
D. 13.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"ans... | D |
question_0037 | (Spatial Construction) Instance/Scene Attribute | [
"question_0037/image_content_0.png",
"question_0037/image_content_1.png"
] | <image><image>Based on the video, which has a higher horizontal height, this potted plant or this piano? | [
"The potted plant and piano have approximately the same horizontal height",
"The piano has a higher horizontal height",
"The potted plant has a higher horizontal height",
"Unable to determine"
] | [
[
"question_0037/3rscan_0006_2/00:00.00_frame_1.00_0.jpg",
"question_0037/3rscan_0006_2/00:01.00_frame_1.00_1.jpg",
"question_0037/3rscan_0006_2/00:02.00_frame_1.00_2.jpg",
"question_0037/3rscan_0006_2/00:03.00_frame_1.00_3.jpg",
"question_0037/3rscan_0006_2/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1,
"end": 16,
"path": "question_0037/3rscan_0006_2.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image><image>Based on the video, which has a higher horizontal height, this potted plant or this piano? | Please select the correct answer from the following options:
A. The potted plant and piano have approximately the same horizontal height;
B. The piano has a higher horizontal height;
C. The potted plant has a higher horizontal height;
D. Unable to determine.
. You need to choose one letter from A,B,C,D as your answer... | A |
question_0038 | (Spatial Construction) Instance/Scene Attribute | [] | How many separate room areas appear in your field of view in this video taken inside a building? | [
"5",
"3",
"4",
"2"
] | [
[
"question_0038/dl3dv-10k_0078_0/00:06.00_frame_1.00_6.jpg",
"question_0038/dl3dv-10k_0078_0/00:07.00_frame_1.00_7.jpg",
"question_0038/dl3dv-10k_0078_0/00:08.00_frame_1.00_8.jpg",
"question_0038/dl3dv-10k_0078_0/00:09.00_frame_1.00_9.jpg",
"question_0038/dl3dv-10k_0078_0/00:10.00_frame_1.00_10... | [
{
"base_fps": 1,
"end": 33,
"path": "question_0038/dl3dv-10k_0078_0.mp4",
"start": 6
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: How many separate room areas appear in your field of view in this video taken inside a building? | Please select the correct answer from the following options:
A. 5;
B. 3;
C. 4;
D. 2.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | A |
question_0039 | (Spatial Construction) Instance/Scene Attribute | [
"question_0039/image_content_0.png",
"question_0039/image_content_1.png"
] | Based on the video information, when you want to move from the position indicated by the red dot to the position indicated by the green dot (<image><image>), which represent locations on the ground, how many possible paths are there without crossing objects? How many? | [
"3",
"6",
"4",
"1"
] | [
[
"question_0039/scannet_fix_0048_0/00:00.00_frame_1.00_0.jpg",
"question_0039/scannet_fix_0048_0/00:01.00_frame_1.00_1.jpg",
"question_0039/scannet_fix_0048_0/00:02.00_frame_1.00_2.jpg",
"question_0039/scannet_fix_0048_0/00:03.00_frame_1.00_3.jpg",
"question_0039/scannet_fix_0048_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 23,
"path": "question_0039/scannet_fix_0048_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Based on the video information, when you want to move from the position indicated by the red dot to the position indicated by the green dot (<image><image>), which represent locations on the ground, how ... | Please select the correct answer from the following options:
A. 3;
B. 6;
C. 4;
D. 1.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | C |
question_0040 | (Spatial Construction) Instance/Scene Attribute | [
"question_0040/image_content_1.png"
] | Based on the information from the video, when you are at these two positions, can you reach the location of the object shown in the image <image> without crossing obstacles? Position 1: Entrance; Position 2: Location of six bottles of beverage | [
"Position 1 is possible, Position 2 is not possible",
"Position 2 is possible, Position 1 is not possible",
"Neither position is possible",
"Both positions are possible"
] | [
[
"question_0040/scannet_fix_0049_0/00:00.00_frame_1.00_0.jpg",
"question_0040/scannet_fix_0049_0/00:01.00_frame_1.00_1.jpg",
"question_0040/scannet_fix_0049_0/00:02.00_frame_1.00_2.jpg",
"question_0040/scannet_fix_0049_0/00:03.00_frame_1.00_3.jpg",
"question_0040/scannet_fix_0049_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 49,
"path": "question_0040/scannet_fix_0049_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Based on the information from the video, when you are at these two positions, can you reach the location of the object shown in the image <image> without crossing obstacles? Position 1: Entrance; Positio... | Please select the correct answer from the following options:
A. Position 1 is possible, Position 2 is not possible;
B. Position 2 is possible, Position 1 is not possible;
C. Neither position is possible;
D. Both positions are possible.
. You need to choose one letter from A,B,C,D as your answer and provide your detai... | D |
question_0041 | (Spatial Construction) Instance/Scene Attribute | [] | How many pillows are there in this room? | [
"13",
"11",
"12",
"14"
] | [
[
"question_0041/arkitscenes_fix_0002_0/00:00.00_frame_1.00_0.jpg",
"question_0041/arkitscenes_fix_0002_0/00:01.00_frame_1.00_1.jpg",
"question_0041/arkitscenes_fix_0002_0/00:02.00_frame_1.00_2.jpg",
"question_0041/arkitscenes_fix_0002_0/00:03.00_frame_1.00_3.jpg",
"question_0041/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 60,
"path": "question_0041/arkitscenes_fix_0002_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: How many pillows are there in this room? | Please select the correct answer from the following options:
A. 13;
B. 11;
C. 12;
D. 14.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"ans... | D |
question_0042 | (Spatial Construction) Instance/Scene Attribute | [] | This video was shot inside an office. How many people can this office accommodate to sit at most? | [
"8",
"7",
"6",
"9"
] | [
[
"question_0042/scannet_fix_0041_0/00:00.00_frame_1.00_0.jpg",
"question_0042/scannet_fix_0041_0/00:01.00_frame_1.00_1.jpg",
"question_0042/scannet_fix_0041_0/00:02.00_frame_1.00_2.jpg",
"question_0042/scannet_fix_0041_0/00:03.00_frame_1.00_3.jpg",
"question_0042/scannet_fix_0041_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 26,
"path": "question_0042/scannet_fix_0041_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video was shot inside an office. How many people can this office accommodate to sit at most? | Please select the correct answer from the following options:
A. 8;
B. 7;
C. 6;
D. 9.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | D |
question_0043 | (Spatial Construction) Instance/Scene Attribute | [
"question_0043/image_content_1.png"
] | Which has a higher horizontal height, the clothing shown in this image <image> or the TV stand base? | [
"Unable to determine",
"Clothing is higher",
"Same height",
"TV stand base is higher"
] | [
[
"question_0043/arkitscenes_fix_0003_0/00:00.00_frame_1.00_0.jpg",
"question_0043/arkitscenes_fix_0003_0/00:01.00_frame_1.00_1.jpg",
"question_0043/arkitscenes_fix_0003_0/00:02.00_frame_1.00_2.jpg",
"question_0043/arkitscenes_fix_0003_0/00:03.00_frame_1.00_3.jpg",
"question_0043/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 33,
"path": "question_0043/arkitscenes_fix_0003_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Which has a higher horizontal height, the clothing shown in this image <image> or the TV stand base? | Please select the correct answer from the following options:
A. Unable to determine;
B. Clothing is higher;
C. Same height;
D. TV stand base is higher.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). On... | B |
question_0044 | (Spatial Construction) Instance/Scene Attribute | [] | How many computer screens are in this room, and how many of them are turned on? | [
"There are 5, all turned on",
"There are 6, with 5 turned on",
"There are 6, with 4 turned on",
"There are 5, with 4 turned on"
] | [
[
"question_0044/scannet_fix_0003_0/00:00.00_frame_1.00_0.jpg",
"question_0044/scannet_fix_0003_0/00:01.00_frame_1.00_1.jpg",
"question_0044/scannet_fix_0003_0/00:02.00_frame_1.00_2.jpg",
"question_0044/scannet_fix_0003_0/00:03.00_frame_1.00_3.jpg",
"question_0044/scannet_fix_0003_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 40,
"path": "question_0044/scannet_fix_0003_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: How many computer screens are in this room, and how many of them are turned on? | Please select the correct answer from the following options:
A. There are 5, all turned on;
B. There are 6, with 5 turned on;
C. There are 6, with 4 turned on;
D. There are 5, with 4 turned on.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what infor... | B |
question_0045 | (Spatial Construction) Instance/Scene Attribute | [] | Is the horizontal level of the golden fish in the water tank now higher, or is the bottom of the TV higher? | [
"Lack of evidence, unable to determine",
"The horizontal level of the bottom of the TV is higher",
"Both are at the same height",
"The horizontal level of the golden fish is higher"
] | [
[
"question_0045/arkitscenes_fix_0052_0/00:26.00_frame_1.00_26.jpg",
"question_0045/arkitscenes_fix_0052_0/00:27.00_frame_1.00_27.jpg",
"question_0045/arkitscenes_fix_0052_0/00:28.00_frame_1.00_28.jpg",
"question_0045/arkitscenes_fix_0052_0/00:29.00_frame_1.00_29.jpg",
"question_0045/arkitscenes... | [
{
"base_fps": 1,
"end": 65,
"path": "question_0045/arkitscenes_fix_0052_0.mp4",
"start": 26
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Is the horizontal level of the golden fish in the water tank now higher, or is the bottom of the TV higher? | Please select the correct answer from the following options:
A. Lack of evidence, unable to determine;
B. The horizontal level of the bottom of the TV is higher;
C. Both are at the same height;
D. The horizontal level of the golden fish is higher.
. You need to choose one letter from A,B,C,D as your answer and provid... | D |
question_0046 | (Spatial Construction) Instance/Scene Attribute | [] | On the street shown in the video, how many lights and how many green square covers can be seen? | [
"3 lights, 4 green square covers",
"3 lights, 5 green square covers",
"2 lights, 5 green square covers",
"2 lights, 4 green square covers"
] | [
[
"question_0046/dl3dv-10k_0032_0/00:25.00_frame_1.00_25.jpg",
"question_0046/dl3dv-10k_0032_0/00:26.00_frame_1.00_26.jpg",
"question_0046/dl3dv-10k_0032_0/00:27.00_frame_1.00_27.jpg",
"question_0046/dl3dv-10k_0032_0/00:28.00_frame_1.00_28.jpg",
"question_0046/dl3dv-10k_0032_0/00:29.00_frame_1.0... | [
{
"base_fps": 1,
"end": 65,
"path": "question_0046/dl3dv-10k_0032_0.mp4",
"start": 25
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: On the street shown in the video, how many lights and how many green square covers can be seen? | Please select the correct answer from the following options:
A. 3 lights, 4 green square covers;
B. 3 lights, 5 green square covers;
C. 2 lights, 5 green square covers;
D. 2 lights, 4 green square covers.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on... | B |
question_0047 | (Spatial Construction) Instance/Scene Attribute | [
"question_0047/image_content_0.png"
] | <image> For the black car shown in this image, to leave this square parking area, how many accessible routes are there in total? | [
"3",
"4",
"5",
"2"
] | [
[
"question_0047/dl3dv_0000_0/00:00.00_frame_1.00_0.jpg",
"question_0047/dl3dv_0000_0/00:01.00_frame_1.00_1.jpg",
"question_0047/dl3dv_0000_0/00:02.00_frame_1.00_2.jpg",
"question_0047/dl3dv_0000_0/00:03.00_frame_1.00_3.jpg",
"question_0047/dl3dv_0000_0/00:04.00_frame_1.00_4.jpg",
"question_... | [
{
"base_fps": 1,
"end": 39,
"path": "question_0047/dl3dv_0000_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image> For the black car shown in this image, to leave this square parking area, how many accessible routes are there in total? | Please select the correct answer from the following options:
A. 3;
B. 4;
C. 5;
D. 2.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0048 | (Spatial Construction) Instance/Scene Attribute | [] | For the central location surrounded by these buildings, how many paths can lead to the outside? | [
"4",
"5",
"3",
"2"
] | [
[
"question_0048/dtu_0007_0/00:00.00_frame_0.50_0.jpg",
"question_0048/dtu_0007_0/00:00.50_frame_0.50_1.jpg",
"question_0048/dtu_0007_0/00:01.00_frame_0.50_2.jpg",
"question_0048/dtu_0007_0/00:01.50_frame_0.50_3.jpg",
"question_0048/dtu_0007_0/00:02.00_frame_0.50_4.jpg",
"question_0048/dtu_0... | [
{
"base_fps": 2,
"end": 24,
"path": "question_0048/dtu_0007_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: For the central location surrounded by these buildings, how many paths can lead to the outside? | Please select the correct answer from the following options:
A. 4;
B. 5;
C. 3;
D. 2.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0049 | (Spatial Construction) Instance/Scene Attribute | [] | Carefully observe these buildings, how many chimneys can be seen? | [
"7",
"5",
"6",
"4"
] | [
[
"question_0049/dtu_0001_0/00:00.00_frame_0.50_0.jpg",
"question_0049/dtu_0001_0/00:00.50_frame_0.50_1.jpg",
"question_0049/dtu_0001_0/00:01.00_frame_0.50_2.jpg",
"question_0049/dtu_0001_0/00:01.50_frame_0.50_3.jpg",
"question_0049/dtu_0001_0/00:02.00_frame_0.50_4.jpg",
"question_0049/dtu_0... | [
{
"base_fps": 2,
"end": 24,
"path": "question_0049/dtu_0001_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the attributes of instances or scenes. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Carefully observe these buildings, how many chimneys can be seen? | Please select the correct answer from the following options:
A. 7;
B. 5;
C. 6;
D. 4.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0050 | Prediction | [
"question_0050/image_content_0.png"
] | It is known that this room is equipped with a kitchen and a bathroom. Assuming you are in the position shown in the image above <image>, if you walk into the room on the left, open the door inside, and keep going straight, which area are you most likely to reach? | [
"bathroom",
"outside",
"none of the above",
"kitchen"
] | [
[
"question_0050/room3d_0167_0/00:00.00_frame_1.00_0.jpg",
"question_0050/room3d_0167_0/00:01.00_frame_1.00_1.jpg",
"question_0050/room3d_0167_0/00:02.00_frame_1.00_2.jpg",
"question_0050/room3d_0167_0/00:03.00_frame_1.00_3.jpg",
"question_0050/room3d_0167_0/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1.000100010001,
"end": 159.984,
"path": "question_0050/room3d_0167_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to simulate or predict outcomes based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: It is known that this room is equipped with a kitchen and a bathroom. Assuming you are in the position shown in the image above <image>, if you walk into the room on the left, open the door inside, and k... | Please select the correct answer from the following options:
A. bathroom;
B. outside;
C. none of the above;
D. kitchen.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer an... | D |
question_0051 | (Motion Understanding) Interactive Motion | [] | In the video, Bob is wiping and arranging the utensils. During this process, how many utensils were successfully placed, and moved a distance greater than the movement of the wooden spoon? | [
"6",
"3",
"5",
"4"
] | [
[
"question_0051/epic-kitchen_0405_0/00:10.00_frame_0.12_80.jpg",
"question_0051/epic-kitchen_0405_0/00:10.12_frame_0.12_81.jpg",
"question_0051/epic-kitchen_0405_0/00:10.25_frame_0.12_82.jpg",
"question_0051/epic-kitchen_0405_0/00:10.38_frame_0.12_83.jpg",
"question_0051/epic-kitchen_0405_0/00:... | [
{
"base_fps": 8,
"end": 34.375,
"path": "question_0051/epic-kitchen_0405_0.mp4",
"start": 10
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: In the video, Bob is wiping and arranging the utensils. During this process, how many utensils were successfully placed, and moved a distance greater than the movement of the wooden spoon? | Please select the correct answer from the following options:
A. 6;
B. 3;
C. 5;
D. 4.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | C |
question_0052 | (Motion Understanding) Interactive Motion | [] | I am organizing objects. What is the order of movement distances for the following objects during this period: cloth, brown rice spoon, black spoon? (Measured by the distance from the starting point to the end point) | [
"Cloth is greater than brown rice spoon greater than black spoon",
"Brown rice spoon is greater than cloth greater than black spoon",
"Black spoon is greater than cloth greater than brown rice spoon",
"Cloth is greater than black spoon greater than brown rice spoon"
] | [
[
"question_0052/epic-kitchen_0405_0/00:00.00_frame_0.12_0.jpg",
"question_0052/epic-kitchen_0405_0/00:00.12_frame_0.12_1.jpg",
"question_0052/epic-kitchen_0405_0/00:00.25_frame_0.12_2.jpg",
"question_0052/epic-kitchen_0405_0/00:00.38_frame_0.12_3.jpg",
"question_0052/epic-kitchen_0405_0/00:00.5... | [
{
"base_fps": 8,
"end": 25,
"path": "question_0052/epic-kitchen_0405_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am organizing objects. What is the order of movement distances for the following objects during this period: cloth, brown rice spoon, black spoon? (Measured by the distance from the starting point to t... | Please select the correct answer from the following options:
A. Cloth is greater than brown rice spoon greater than black spoon;
B. Brown rice spoon is greater than cloth greater than black spoon;
C. Black spoon is greater than cloth greater than brown rice spoon;
D. Cloth is greater than black spoon greater than bro... | B |
question_0053 | (Motion Understanding) Interactive Motion | [] | I am operating a robotic arm to install a light bulb. During this process, how many times did the robotic arm rotate the light bulb? | [
"6",
"9",
"7",
"8"
] | [
[
"question_0053/rh20t_0050_2/00:50.00_frame_0.25_200.jpg",
"question_0053/rh20t_0050_2/00:50.25_frame_0.25_201.jpg",
"question_0053/rh20t_0050_2/00:50.50_frame_0.25_202.jpg",
"question_0053/rh20t_0050_2/00:50.75_frame_0.25_203.jpg",
"question_0053/rh20t_0050_2/00:51.00_frame_0.25_204.jpg",
... | [
{
"base_fps": 4,
"end": 80.25,
"path": "question_0053/rh20t_0050_2.mp4",
"start": 50
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am operating a robotic arm to install a light bulb. During this process, how many times did the robotic arm rotate the light bulb? | Please select the correct answer from the following options:
A. 6;
B. 9;
C. 7;
D. 8.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | C |
question_0054 | (Motion Understanding) Interactive Motion | [] | The video shows a person in the process of cutting potatoes. When he is about to cut the fourth piece of potato, by approximately how many degrees has the entire potato been rotated (counting from the 0-second mark)? | [
"180",
"270",
"30",
"90"
] | [
[
"question_0054/egoexolearn_0021_0/00:00.00_frame_0.25_0.jpg",
"question_0054/egoexolearn_0021_0/00:00.25_frame_0.25_1.jpg",
"question_0054/egoexolearn_0021_0/00:00.50_frame_0.25_2.jpg",
"question_0054/egoexolearn_0021_0/00:00.75_frame_0.25_3.jpg",
"question_0054/egoexolearn_0021_0/00:01.00_fra... | [
{
"base_fps": 4,
"end": 22.5,
"path": "question_0054/egoexolearn_0021_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video shows a person in the process of cutting potatoes. When he is about to cut the fourth piece of potato, by approximately how many degrees has the entire potato been rotated (counting from the 0-... | Please select the correct answer from the following options:
A. 180;
B. 270;
C. 30;
D. 90.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"a... | A |
question_0055 | (Motion Understanding) Interactive Motion | [] | During this period, how many times did the cook in the video obviously change the position of this piece of pork in the pan in total? | [
"5",
"3",
"8",
"6"
] | [
[
"question_0055/egoexolearn_0086_0/00:00.00_frame_0.25_0.jpg",
"question_0055/egoexolearn_0086_0/00:00.25_frame_0.25_1.jpg",
"question_0055/egoexolearn_0086_0/00:00.50_frame_0.25_2.jpg",
"question_0055/egoexolearn_0086_0/00:00.75_frame_0.25_3.jpg",
"question_0055/egoexolearn_0086_0/00:01.00_fra... | [
{
"base_fps": 4,
"end": 25,
"path": "question_0055/egoexolearn_0086_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: During this period, how many times did the cook in the video obviously change the position of this piece of pork in the pan in total? | Please select the correct answer from the following options:
A. 5;
B. 3;
C. 8;
D. 6.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | D |
question_0056 | (Motion Understanding) Interactive Motion | [
"question_0056/image_content_0.png"
] | <image> The video shows me assembling these wooden blocks. During this time, I took the wooden blocks at the positions of the red dot and green dot as shown in the image. At the end of the video, where are they respectively? | [
"The red dot wooden block is in hand at the end; the green dot wooden block is the bottom one of the three assembled blocks in view.",
"The green dot wooden block is the leftmost one of the three assembled blocks in view at the end; the red dot wooden block is the bottom one.",
"The green dot wooden block is in... | [
[
"question_0056/ego4d_0081_0/01:53.50_frame_0.50_227.jpg",
"question_0056/ego4d_0081_0/01:54.00_frame_0.50_228.jpg",
"question_0056/ego4d_0081_0/01:54.50_frame_0.50_229.jpg",
"question_0056/ego4d_0081_0/01:55.00_frame_0.50_230.jpg",
"question_0056/ego4d_0081_0/01:55.50_frame_0.50_231.jpg",
... | [
{
"base_fps": 2,
"end": 213.5,
"path": "question_0056/ego4d_0081_0.mp4",
"start": 113.5
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image> The video shows me assembling these wooden blocks. During this time, I took the wooden blocks at the positions of the red dot and green dot as shown in the image. At the end of the video, where a... | Please select the correct answer from the following options:
A. The red dot wooden block is in hand at the end; the green dot wooden block is the bottom one of the three assembled blocks in view.;
B. The green dot wooden block is the leftmost one of the three assembled blocks in view at the end; the red dot wooden bl... | A |
question_0057 | (Motion Understanding) Interactive Motion | [] | The video captures the actions of a person, during which objects are moved. How are these objects moved? (Considering the direction from the staircase to the camera as north) | [
"First moved north, then south, and finally east",
"First moved north, then south, and finally west",
"First moved south, then north, and finally east",
"First moved south, then north, and finally west"
] | [
[
"question_0057/charades_0011_0/00:00.00_frame_0.25_0.jpg",
"question_0057/charades_0011_0/00:00.25_frame_0.25_1.jpg",
"question_0057/charades_0011_0/00:00.50_frame_0.25_2.jpg",
"question_0057/charades_0011_0/00:00.75_frame_0.25_3.jpg",
"question_0057/charades_0011_0/00:01.00_frame_0.25_4.jpg",... | [
{
"base_fps": 4,
"end": 23.5,
"path": "question_0057/charades_0011_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video captures the actions of a person, during which objects are moved. How are these objects moved? (Considering the direction from the staircase to the camera as north) | Please select the correct answer from the following options:
A. First moved north, then south, and finally east;
B. First moved north, then south, and finally west;
C. First moved south, then north, and finally east;
D. First moved south, then north, and finally west.
. You need to choose one letter from A,B,C,D as y... | A |
question_0058 | (Motion Understanding) Interactive Motion | [] | The video shows a person tidying up the room. During this process, how many objects were significantly moved, and how many objects ended up noticeably more to the left in the frame compared to their initial position? | [
"9 objects were significantly moved, 3 objects ended up noticeably more to the left in the frame compared to their initial position.",
"9 objects were significantly moved, 4 objects ended up noticeably more to the left in the frame compared to their initial position.",
"10 objects were significantly moved, 3 ob... | [
[
"question_0058/charades_0055_0/00:00.00_frame_0.25_0.jpg",
"question_0058/charades_0055_0/00:00.25_frame_0.25_1.jpg",
"question_0058/charades_0055_0/00:00.50_frame_0.25_2.jpg",
"question_0058/charades_0055_0/00:00.75_frame_0.25_3.jpg",
"question_0058/charades_0055_0/00:01.00_frame_0.25_4.jpg",... | [
{
"base_fps": 4,
"end": 40.5,
"path": "question_0058/charades_0055_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video shows a person tidying up the room. During this process, how many objects were significantly moved, and how many objects ended up noticeably more to the left in the frame compared to their init... | Please select the correct answer from the following options:
A. 9 objects were significantly moved, 3 objects ended up noticeably more to the left in the frame compared to their initial position.;
B. 9 objects were significantly moved, 4 objects ended up noticeably more to the left in the frame compared to their init... | E |
question_0059 | (Motion Understanding) Interactive Motion | [
"question_0059/image_content_0.png",
"question_0059/image_content_1.png",
"question_0059/image_content_2.png"
] | How many of the three objects in the video <image><image><image> have noticeably moved away from the painting due to human activity during this period? | [
"3",
"1",
"2",
"0"
] | [
[
"question_0059/charades_0000_0/00:00.00_frame_0.25_0.jpg",
"question_0059/charades_0000_0/00:00.25_frame_0.25_1.jpg",
"question_0059/charades_0000_0/00:00.50_frame_0.25_2.jpg",
"question_0059/charades_0000_0/00:00.75_frame_0.25_3.jpg",
"question_0059/charades_0000_0/00:01.00_frame_0.25_4.jpg",... | [
{
"base_fps": 4,
"end": 31.75,
"path": "question_0059/charades_0000_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: How many of the three objects in the video <image><image><image> have noticeably moved away from the painting due to human activity during this period? | Please select the correct answer from the following options:
A. 3;
B. 1;
C. 2;
D. 0.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0060 | (Cross-Video) Memoery Update | [] | Mark usually likes to record daily arrangements. These are videos shot during two time periods. Compared to the scene shot in the first time period, which description of the changes in the scene in the second time period is correct? | [
"At least two new objects have been added",
"The position of the pillow has changed",
"Someone went out to work or school",
"The position of the chair has changed"
] | [
[
"question_0060/3rscan_0000_0/00:00.00_frame_1.00_0.jpg",
"question_0060/3rscan_0000_0/00:01.00_frame_1.00_1.jpg",
"question_0060/3rscan_0000_0/00:02.00_frame_1.00_2.jpg",
"question_0060/3rscan_0000_0/00:03.00_frame_1.00_3.jpg",
"question_0060/3rscan_0000_0/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1,
"end": 50,
"path": "question_0060/3rscan_0000_0.mp4",
"start": 0
},
{
"base_fps": 1,
"end": 43,
"path": "question_0060/3rscan_0000_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Mark usually likes to record daily arrangements. These are videos shot during two time periods. Compared to the scene shot in the first time period, ... | Please select the correct answer from the following options:
A. At least two new objects have been added;
B. The position of the pillow has changed;
C. Someone went out to work or school;
D. The position of the chair has changed.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed re... | C |
question_0061 | (Cross-Video) Memoery Update | [
"question_0061/image_content_1.png",
"question_0061/image_content_2.png",
"question_0061/image_content_3.png",
"question_0061/image_content_4.png"
] | The owner of this room likes to take pictures of their room's arrangement at regular intervals. During this period, among the objects shown in the images <image><image><image><image>, how many disappeared from the scene and how many are still present but have changed positions? | [
"1 disappeared from the scene, 2 are still present but have changed positions",
"1 disappeared from the scene, 1 is still present but has changed positions",
"None disappeared from the scene, 1 is still present but has changed position",
"2 disappeared from the scene, none are still present but have changed p... | [
[
"question_0061/3rscan_0000_0/00:00.00_frame_1.00_0.jpg",
"question_0061/3rscan_0000_0/00:01.00_frame_1.00_1.jpg",
"question_0061/3rscan_0000_0/00:02.00_frame_1.00_2.jpg",
"question_0061/3rscan_0000_0/00:03.00_frame_1.00_3.jpg",
"question_0061/3rscan_0000_0/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1,
"end": 50,
"path": "question_0061/3rscan_0000_0.mp4",
"start": 0
},
{
"base_fps": 1,
"end": 64,
"path": "question_0061/3rscan_0000_3.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The owner of this room likes to take pictures of their room's arrangement at regular intervals. During this period, among the objects shown in the im... | Please select the correct answer from the following options:
A. 1 disappeared from the scene, 2 are still present but have changed positions;
B. 1 disappeared from the scene, 1 is still present but has changed positions;
C. None disappeared from the scene, 1 is still present but has changed position;
D. 2 disappeared... | A |
question_0062 | (Cross-Video) Memoery Update | [] | The first video shows that I randomly placed the drinks on the ground when leaving the scene. When I returned and shot the second video as shown in the image, what happened to the two buckets of drinks? | [
"Both buckets of drinks disappeared",
"One bucket of drinks disappeared, the other remained in its original position",
"One bucket of drinks disappeared, the other was moved down by someone",
"One bucket of drinks disappeared, the other was moved up by someone"
] | [
[
"question_0062/realworld_0000_0/00:00.00_frame_0.50_0.jpg",
"question_0062/realworld_0000_0/00:00.50_frame_0.50_1.jpg",
"question_0062/realworld_0000_0/00:01.00_frame_0.50_2.jpg",
"question_0062/realworld_0000_0/00:01.50_frame_0.50_3.jpg",
"question_0062/realworld_0000_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 22,
"path": "question_0062/realworld_0000_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 21,
"path": "question_0062/realworld_0000_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The first video shows that I randomly placed the drinks on the ground when leaving the scene. When I returned and shot the second video as shown in t... | Please select the correct answer from the following options:
A. Both buckets of drinks disappeared;
B. One bucket of drinks disappeared, the other remained in its original position;
C. One bucket of drinks disappeared, the other was moved down by someone;
D. One bucket of drinks disappeared, the other was moved up by... | C |
question_0063 | (Cross-Video) Memoery Update | [] | This is Alex shooting the same room at two different times. Based on the video, he captured possible human activities that might have occurred during the time he was not present. Which of the following must have occurred? | [
"Someone used the computer",
"Someone looked through items inside the backpack",
"Someone refilled water at the dispenser",
"Someone browsed through books"
] | [
[
"question_0063/realworld_0016_0/00:00.00_frame_0.50_0.jpg",
"question_0063/realworld_0016_0/00:00.50_frame_0.50_1.jpg",
"question_0063/realworld_0016_0/00:01.00_frame_0.50_2.jpg",
"question_0063/realworld_0016_0/00:01.50_frame_0.50_3.jpg",
"question_0063/realworld_0016_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 43.5,
"path": "question_0063/realworld_0016_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 36,
"path": "question_0063/realworld_0016_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This is Alex shooting the same room at two different times. Based on the video, he captured possible human activities that might have occurred during... | Please select the correct answer from the following options:
A. Someone used the computer;
B. Someone looked through items inside the backpack;
C. Someone refilled water at the dispenser;
D. Someone browsed through books.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning ... | A |
question_0064 | (Cross-Video) Memoery Update | [] | The school security records the parking situation of vehicles during the day and night. Based on these two videos, how many vehicles have left and how many new ones have been added in the row of parking spaces closest to the wall? | [
"2 vehicles have left, 0 new vehicles have been added",
"3 vehicles have left, 2 new vehicles have been added",
"2 vehicles have left, 1 new vehicle has been added",
"3 vehicles have left, 1 new vehicle has been added"
] | [
[
"question_0064/realworld_0005_1/00:00.00_frame_0.50_0.jpg",
"question_0064/realworld_0005_1/00:00.50_frame_0.50_1.jpg",
"question_0064/realworld_0005_1/00:01.00_frame_0.50_2.jpg",
"question_0064/realworld_0005_1/00:01.50_frame_0.50_3.jpg",
"question_0064/realworld_0005_1/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 33,
"path": "question_0064/realworld_0005_1.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 37,
"path": "question_0064/realworld_0005_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The school security records the parking situation of vehicles during the day and night. Based on these two videos, how many vehicles have left and ho... | Please select the correct answer from the following options:
A. 2 vehicles have left, 0 new vehicles have been added;
B. 3 vehicles have left, 2 new vehicles have been added;
C. 2 vehicles have left, 1 new vehicle has been added;
D. 3 vehicles have left, 1 new vehicle has been added.
. You need to choose one letter f... | A |
question_0065 | (Cross-Video) Memoery Update | [
"question_0065/image_content_0.png",
"question_0065/image_content_1.png",
"question_0065/image_content_2.png",
"question_0065/image_content_3.png"
] | <image><image><image><image>Two videos recorded the parked vehicles at a certain time during the day and night. Regarding the four conspicuous-colored cars highlighted in the image, which description is correct? | [
"One car left the parking lot and one car changed its position in the parking lot",
"One car left the parking lot",
"No cars moved",
"Two cars changed their positions in the parking lot"
] | [
[
"question_0065/realworld_0005_1/00:00.00_frame_0.50_0.jpg",
"question_0065/realworld_0005_1/00:00.50_frame_0.50_1.jpg",
"question_0065/realworld_0005_1/00:01.00_frame_0.50_2.jpg",
"question_0065/realworld_0005_1/00:01.50_frame_0.50_3.jpg",
"question_0065/realworld_0005_1/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 33,
"path": "question_0065/realworld_0005_1.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 37,
"path": "question_0065/realworld_0005_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image><image><image><image>Two videos recorded the parked vehicles at a certain time during the day and night. Regarding the four conspicuous-colore... | Please select the correct answer from the following options:
A. One car left the parking lot and one car changed its position in the parking lot;
B. One car left the parking lot;
C. No cars moved;
D. Two cars changed their positions in the parking lot.
. You need to choose one letter from A,B,C,D as your answer and p... | C |
question_0066 | (Cross-Video) Memoery Update | [] | Were the drinks I left on the table moved by someone while I was away, as captured in these two different video clips filmed at separate times? How have they been moved? | [
"No one moved the drinks",
"The drinks moved one table position closer to the red square building",
"The drinks moved two table positions away from the red square building",
"The drinks moved one table position away from the red square building"
] | [
[
"question_0066/realworld_0008_0/00:00.00_frame_0.50_0.jpg",
"question_0066/realworld_0008_0/00:00.50_frame_0.50_1.jpg",
"question_0066/realworld_0008_0/00:01.00_frame_0.50_2.jpg",
"question_0066/realworld_0008_0/00:01.50_frame_0.50_3.jpg",
"question_0066/realworld_0008_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 21.5,
"path": "question_0066/realworld_0008_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 17.5,
"path": "question_0066/realworld_0008_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Were the drinks I left on the table moved by someone while I was away, as captured in these two different video clips filmed at separate times? How h... | Please select the correct answer from the following options:
A. No one moved the drinks;
B. The drinks moved one table position closer to the red square building;
C. The drinks moved two table positions away from the red square building;
D. The drinks moved one table position away from the red square building.
. You ... | D |
question_0067 | (Cross-Video) Memoery Update | [
"question_0067/image_content_1.png",
"question_0067/image_content_2.png",
"question_0067/image_content_3.png"
] | These two video segments were taken by our company's cleaning staff at two different times. Please tell, for the three types of chairs shown in the images below, how many types of chairs have changed position? <image><image><image> | [
"0",
"2",
"3",
"1"
] | [
[
"question_0067/realworld_0025_0/00:00.00_frame_0.50_0.jpg",
"question_0067/realworld_0025_0/00:00.50_frame_0.50_1.jpg",
"question_0067/realworld_0025_0/00:01.00_frame_0.50_2.jpg",
"question_0067/realworld_0025_0/00:01.50_frame_0.50_3.jpg",
"question_0067/realworld_0025_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 24.5,
"path": "question_0067/realworld_0025_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 19.5,
"path": "question_0067/realworld_0025_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: These two video segments were taken by our company's cleaning staff at two different times. Please tell, for the three types of chairs shown in the i... | Please select the correct answer from the following options:
A. 0;
B. 2;
C. 3;
D. 1.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0068 | (Cross-Video) Memoery Update | [
"question_0068/image_content_1.png",
"question_0068/image_content_2.png"
] | These two videos were shot in sequence, before leaving home and after returning home, respectively. For the objects outlined in red in the picture, which description is correct? <image><image> Hint: The TV is on the north side of this room and has not moved. | [
"Both the pillow and the green sign moved to the east",
"The positions of the pillow and the green sign have not changed",
"The position of the green sign has not changed, the pillow moved to the west",
"The green sign moved to the east, the position of the pillow has not changed",
"The green sign moved to ... | [
[
"question_0068/3rscan_0003_1/00:00.00_frame_1.00_0.jpg",
"question_0068/3rscan_0003_1/00:01.00_frame_1.00_1.jpg",
"question_0068/3rscan_0003_1/00:02.00_frame_1.00_2.jpg",
"question_0068/3rscan_0003_1/00:03.00_frame_1.00_3.jpg",
"question_0068/3rscan_0003_1/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1,
"end": 19,
"path": "question_0068/3rscan_0003_1.mp4",
"start": 0
},
{
"base_fps": 1,
"end": 49,
"path": "question_0068/3rscan_0003_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: These two videos were shot in sequence, before leaving home and after returning home, respectively. For the objects outlined in red in the picture, w... | Please select the correct answer from the following options:
A. Both the pillow and the green sign moved to the east;
B. The positions of the pillow and the green sign have not changed;
C. The position of the green sign has not changed, the pillow moved to the west;
D. The green sign moved to the east, the position o... | A |
question_0069 | (Cross-Video) Memoery Update | [] | I initially recorded the first video, followed by the second video. How many additional white tables can you determine exist through the second video? | [
"1",
"0",
"2",
"3"
] | [
[
"question_0069/3rscan_0002_1/00:00.00_frame_1.00_0.jpg",
"question_0069/3rscan_0002_1/00:01.00_frame_1.00_1.jpg",
"question_0069/3rscan_0002_1/00:02.00_frame_1.00_2.jpg",
"question_0069/3rscan_0002_1/00:03.00_frame_1.00_3.jpg",
"question_0069/3rscan_0002_1/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1,
"end": 26,
"path": "question_0069/3rscan_0002_1.mp4",
"start": 0
},
{
"base_fps": 1,
"end": 34,
"path": "question_0069/3rscan_0002_2.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I initially recorded the first video, followed by the second video. How many additional white tables can you determine exist through the second video... | Please select the correct answer from the following options:
A. 1;
B. 0;
C. 2;
D. 3.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | D |
question_0070 | (Motion Understanding) Interactive Motion | [] | I am maneuvering a robotic arm to grab things. How has the number of stationery items in the cup far from the lens changed during this period? | [
"Increased by two",
"Decreased by two",
"Increased by one",
"Decreased by one"
] | [
[
"question_0070/oxe_0115_1/00:00.00_frame_0.25_0.jpg",
"question_0070/oxe_0115_1/00:00.25_frame_0.25_1.jpg",
"question_0070/oxe_0115_1/00:00.50_frame_0.25_2.jpg",
"question_0070/oxe_0115_1/00:00.75_frame_0.25_3.jpg",
"question_0070/oxe_0115_1/00:01.00_frame_0.25_4.jpg",
"question_0070/oxe_0... | [
{
"base_fps": 4,
"end": 50,
"path": "question_0070/oxe_0115_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am maneuvering a robotic arm to grab things. How has the number of stationery items in the cup far from the lens changed during this period? | Please select the correct answer from the following options:
A. Increased by two;
B. Decreased by two;
C. Increased by one;
D. Decreased by one.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only prov... | A |
question_0071 | (Motion Understanding) Interactive Motion | [] | In the video, the main character is organizing the cabinet. When does she start moving items inside the cabinet for the fourth time? | [
"18.5s",
"13.5s",
"15s",
"20s"
] | [
[
"question_0071/charades_0036_0/00:00.00_frame_0.25_0.jpg",
"question_0071/charades_0036_0/00:00.25_frame_0.25_1.jpg",
"question_0071/charades_0036_0/00:00.50_frame_0.25_2.jpg",
"question_0071/charades_0036_0/00:00.75_frame_0.25_3.jpg",
"question_0071/charades_0036_0/00:01.00_frame_0.25_4.jpg",... | [
{
"base_fps": 4,
"end": 29.75,
"path": "question_0071/charades_0036_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: In the video, the main character is organizing the cabinet. When does she start moving items inside the cabinet for the fourth time? | Please select the correct answer from the following options:
A. 18.5s;
B. 13.5s;
C. 15s;
D. 20s.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON forma... | A |
question_0072 | (Motion Understanding) Interactive Motion | [] | At 16 seconds, which card does the person closest to the camera hand to the girl, counted from the top down? | [
"Second card on her left hand side",
"First card on her right hand side",
"First card on her left hand side",
"Second card on her right hand side"
] | [
[
"question_0072/ego4d-highfps_0060_0/00:00.00_frame_0.12_0.jpg",
"question_0072/ego4d-highfps_0060_0/00:00.12_frame_0.12_1.jpg",
"question_0072/ego4d-highfps_0060_0/00:00.25_frame_0.12_2.jpg",
"question_0072/ego4d-highfps_0060_0/00:00.38_frame_0.12_3.jpg",
"question_0072/ego4d-highfps_0060_0/00... | [
{
"base_fps": 8,
"end": 16,
"path": "question_0072/ego4d-highfps_0060_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: At 16 seconds, which card does the person closest to the camera hand to the girl, counted from the top down? | Please select the correct answer from the following options:
A. Second card on her left hand side;
B. First card on her right hand side;
C. First card on her left hand side;
D. Second card on her right hand side.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (... | D |
question_0073 | (Motion Understanding) Interactive Motion | [] | The video recorded my activities in the room. Based on the video, how many seconds in total are my hands empty? | [
"25s",
"10s",
"15s",
"20s"
] | [
[
"question_0073/ego4d_0020_0/00:00.00_frame_0.50_0.jpg",
"question_0073/ego4d_0020_0/00:00.50_frame_0.50_1.jpg",
"question_0073/ego4d_0020_0/00:01.00_frame_0.50_2.jpg",
"question_0073/ego4d_0020_0/00:01.50_frame_0.50_3.jpg",
"question_0073/ego4d_0020_0/00:02.00_frame_0.50_4.jpg",
"question_... | [
{
"base_fps": 2,
"end": 56,
"path": "question_0073/ego4d_0020_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video recorded my activities in the room. Based on the video, how many seconds in total are my hands empty? | Please select the correct answer from the following options:
A. 25s;
B. 10s;
C. 15s;
D. 20s.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {... | C |
question_0074 | (Motion Understanding) Interactive Motion | [] | The video shows a painter painting. During which time period did his brush cover the largest area? | [
"40-48 seconds",
"0-8 seconds",
"20-28 seconds",
"30-38 seconds"
] | [
[
"question_0074/ego4d_0007_0/00:00.00_frame_0.50_0.jpg",
"question_0074/ego4d_0007_0/00:00.50_frame_0.50_1.jpg",
"question_0074/ego4d_0007_0/00:01.00_frame_0.50_2.jpg",
"question_0074/ego4d_0007_0/00:01.50_frame_0.50_3.jpg",
"question_0074/ego4d_0007_0/00:02.00_frame_0.50_4.jpg",
"question_... | [
{
"base_fps": 2,
"end": 75,
"path": "question_0074/ego4d_0007_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The video shows a painter painting. During which time period did his brush cover the largest area? | Please select the correct answer from the following options:
A. 40-48 seconds;
B. 0-8 seconds;
C. 20-28 seconds;
D. 30-38 seconds.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your outpu... | D |
question_0075 | (Motion Understanding) Interactive Motion | [
"question_0075/image_content_0.png"
] | At the beginning of the video, where is the card shown in this image <image>? | [
"Second card in the deck",
"In the discard pile",
"In the main perspective's hand",
"At the top of the deck"
] | [
[
"question_0075/ego4d-highfps_0060_0/01:56.25_frame_0.12_930.jpg",
"question_0075/ego4d-highfps_0060_0/01:56.38_frame_0.12_931.jpg",
"question_0075/ego4d-highfps_0060_0/01:56.50_frame_0.12_932.jpg",
"question_0075/ego4d-highfps_0060_0/01:56.62_frame_0.12_933.jpg",
"question_0075/ego4d-highfps_0... | [
{
"base_fps": 8,
"end": 128.75,
"path": "question_0075/ego4d-highfps_0060_0.mp4",
"start": 116.25
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: At the beginning of the video, where is the card shown in this image <image>? | Please select the correct answer from the following options:
A. Second card in the deck;
B. In the discard pile;
C. In the main perspective's hand;
D. At the top of the deck.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived ... | A |
question_0076 | (Motion Understanding) Interactive Motion | [] | The protagonist in the video is practicing writing. When does she begin practicing the second line of the second sheet of writing paper? | [
"45s",
"30s",
"1m15s",
"1m"
] | [
[
"question_0076/ego4d_0062_0/00:00.00_frame_0.50_0.jpg",
"question_0076/ego4d_0062_0/00:00.50_frame_0.50_1.jpg",
"question_0076/ego4d_0062_0/00:01.00_frame_0.50_2.jpg",
"question_0076/ego4d_0062_0/00:01.50_frame_0.50_3.jpg",
"question_0076/ego4d_0062_0/00:02.00_frame_0.50_4.jpg",
"question_... | [
{
"base_fps": 2,
"end": 100,
"path": "question_0076/ego4d_0062_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The protagonist in the video is practicing writing. When does she begin practicing the second line of the second sheet of writing paper? | Please select the correct answer from the following options:
A. 45s;
B. 30s;
C. 1m15s;
D. 1m.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: ... | C |
question_0077 | (Motion Understanding) Interactive Motion | [] | When does the fourth bamboo cut occur in the video? | [
"Around 15 seconds",
"Around 5 seconds",
"Around 10 seconds",
"Around 20 seconds"
] | [
[
"question_0077/ego4d-highfps_0033_0/00:00.00_frame_0.12_0.jpg",
"question_0077/ego4d-highfps_0033_0/00:00.12_frame_0.12_1.jpg",
"question_0077/ego4d-highfps_0033_0/00:00.25_frame_0.12_2.jpg",
"question_0077/ego4d-highfps_0033_0/00:00.38_frame_0.12_3.jpg",
"question_0077/ego4d-highfps_0033_0/00... | [
{
"base_fps": 8,
"end": 20,
"path": "question_0077/ego4d-highfps_0033_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: When does the fourth bamboo cut occur in the video? | Please select the correct answer from the following options:
A. Around 15 seconds;
B. Around 5 seconds;
C. Around 10 seconds;
D. Around 20 seconds.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only p... | C |
question_0078 | (Motion Understanding) Interactive Motion | [] | Tom is knitting clothes. During this time, how many times did the yarn on his left hand wrap around the tool in total? | [
"5",
"7",
"6",
"4"
] | [
[
"question_0078/ego4d-highfps_0021_0/00:06.25_frame_0.12_50.jpg",
"question_0078/ego4d-highfps_0021_0/00:06.38_frame_0.12_51.jpg",
"question_0078/ego4d-highfps_0021_0/00:06.50_frame_0.12_52.jpg",
"question_0078/ego4d-highfps_0021_0/00:06.62_frame_0.12_53.jpg",
"question_0078/ego4d-highfps_0021_... | [
{
"base_fps": 8,
"end": 31.25,
"path": "question_0078/ego4d-highfps_0021_0.mp4",
"start": 6.25
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Tom is knitting clothes. During this time, how many times did the yarn on his left hand wrap around the tool in total? | Please select the correct answer from the following options:
A. 5;
B. 7;
C. 6;
D. 4.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0079 | (Motion Understanding) Interactive Motion | [
"question_0079/image_content_0.png"
] | In the video, the person is teasing the cat with a ribbon, and there is a post in front of him. How many times in total does the ribbon circle around this post (i.e., passing the red dot position behind the post shown in the picture <image>)? When is the first time the ribbon circles the post counterclockwise? | [
"2 times, at 36-38 seconds",
"1 time, at 36-38 seconds",
"1 time, at 44-46 seconds",
"2 times, at 44-46 seconds"
] | [
[
"question_0079/ego4d-highfps_0092_0/00:33.75_frame_0.12_270.jpg",
"question_0079/ego4d-highfps_0092_0/00:33.88_frame_0.12_271.jpg",
"question_0079/ego4d-highfps_0092_0/00:34.00_frame_0.12_272.jpg",
"question_0079/ego4d-highfps_0092_0/00:34.12_frame_0.12_273.jpg",
"question_0079/ego4d-highfps_0... | [
{
"base_fps": 8,
"end": 58,
"path": "question_0079/ego4d-highfps_0092_0.mp4",
"start": 33.75
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task focuses on the interactions between instances. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: In the video, the person is teasing the cat with a ribbon, and there is a post in front of him. How many times in total does the ribbon circle around this post (i.e., passing the red dot position behind ... | Please select the correct answer from the following options:
A. 2 times, at 36-38 seconds;
B. 1 time, at 36-38 seconds;
C. 1 time, at 44-46 seconds;
D. 2 times, at 44-46 seconds.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arri... | D |
question_0080 | (Cross-Video) Memoery Update | [] | This is the scene I saw when passing through the stairwell at two different times. What should I do to restore the arrangement of objects in the background to how it was the first time I passed through the stairwell? | [
"No action needed",
"Move the umbrella up three flights of stairs",
"Move the umbrella up one flight of stairs, move the trash can down two flights of stairs",
"Move the trash can down three flights of stairs"
] | [
[
"question_0080/realworld_0010_0/00:00.00_frame_0.50_0.jpg",
"question_0080/realworld_0010_0/00:00.50_frame_0.50_1.jpg",
"question_0080/realworld_0010_0/00:01.00_frame_0.50_2.jpg",
"question_0080/realworld_0010_0/00:01.50_frame_0.50_3.jpg",
"question_0080/realworld_0010_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 28.5,
"path": "question_0080/realworld_0010_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 33,
"path": "question_0080/realworld_0010_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This is the scene I saw when passing through the stairwell at two different times. What should I do to restore the arrangement of objects in the back... | Please select the correct answer from the following options:
A. No action needed;
B. Move the umbrella up three flights of stairs;
C. Move the umbrella up one flight of stairs, move the trash can down two flights of stairs;
D. Move the trash can down three flights of stairs.
. You need to choose one letter from A,B,C... | B |
question_0081 | (Cross-Video) Memoery Update | [] | The first video was recorded by a tourist with a camera, and later he revisited the place and shot the second video. Are the two trees clearly recorded in succession in the second video (at the 5th second and at the last moment) seen before? | [
"Neither of the two has been seen before",
"The first tree has been seen before, the second tree has not been seen before",
"Both of the two have been seen before",
"The first tree has not been seen before, the second tree has been seen before"
] | [
[
"question_0081/realworld_0024_0/00:00.00_frame_0.50_0.jpg",
"question_0081/realworld_0024_0/00:00.50_frame_0.50_1.jpg",
"question_0081/realworld_0024_0/00:01.00_frame_0.50_2.jpg",
"question_0081/realworld_0024_0/00:01.50_frame_0.50_3.jpg",
"question_0081/realworld_0024_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 20.5,
"path": "question_0081/realworld_0024_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 18.5,
"path": "question_0081/realworld_0024_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: The first video was recorded by a tourist with a camera, and later he revisited the place and shot the second video. Are the two trees clearly record... | Please select the correct answer from the following options:
A. Neither of the two has been seen before;
B. The first tree has been seen before, the second tree has not been seen before;
C. Both of the two have been seen before;
D. The first tree has not been seen before, the second tree has been seen before.
. You n... | B |
question_0082 | (Cross-Video) Memoery Update | [] | These two videos were taken by our company's cleaning staff at noon and in the afternoon, and no arrangements were moved during the shooting. Apart from the room with the computer, how many of these rooms were used between noon and afternoon? | [
"0",
"1",
"3",
"2"
] | [
[
"question_0082/realworld_0007_0/00:00.00_frame_0.50_0.jpg",
"question_0082/realworld_0007_0/00:00.50_frame_0.50_1.jpg",
"question_0082/realworld_0007_0/00:01.00_frame_0.50_2.jpg",
"question_0082/realworld_0007_0/00:01.50_frame_0.50_3.jpg",
"question_0082/realworld_0007_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 69,
"path": "question_0082/realworld_0007_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 50,
"path": "question_0082/realworld_0007_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: These two videos were taken by our company's cleaning staff at noon and in the afternoon, and no arrangements were moved during the shooting. Apart f... | Please select the correct answer from the following options:
A. 0;
B. 1;
C. 3;
D. 2.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | C |
question_0083 | (Cross-Video) Memoery Update | [] | I recorded these two videos at two different times. Based on the video content, how many objects can you determine have changed positions at most? (Excluding chairs) | [
"6",
"2",
"3",
"5",
"4",
"7"
] | [
[
"question_0083/realworld_0014_0/00:00.00_frame_0.50_0.jpg",
"question_0083/realworld_0014_0/00:00.50_frame_0.50_1.jpg",
"question_0083/realworld_0014_0/00:01.00_frame_0.50_2.jpg",
"question_0083/realworld_0014_0/00:01.50_frame_0.50_3.jpg",
"question_0083/realworld_0014_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 30.5,
"path": "question_0083/realworld_0014_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 33.5,
"path": "question_0083/realworld_0014_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I recorded these two videos at two different times. Based on the video content, how many objects can you determine have changed positions at most? (E... | Please select the correct answer from the following options:
A. 6;
B. 2;
C. 3;
D. 5;
E. 4;
F. 7.
. You need to choose one letter from A,B,C,D,E,F as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON f... | A |
question_0084 | (Cross-Video) Memoery Update | [] | John currently shares an apartment with another person, and his bed is the one by the window. These two videos were recorded by him before and after his roommate entered and exited the room. Based on this, did John's roommate move anything on John's bed or rearrange anything in the bathroom? | [
"He did not move anything in either place",
"He moved things in both places",
"He only moved things on John's bed",
"He only moved things in the bathroom"
] | [
[
"question_0084/realworld_0014_0/00:00.00_frame_0.50_0.jpg",
"question_0084/realworld_0014_0/00:00.50_frame_0.50_1.jpg",
"question_0084/realworld_0014_0/00:01.00_frame_0.50_2.jpg",
"question_0084/realworld_0014_0/00:01.50_frame_0.50_3.jpg",
"question_0084/realworld_0014_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 30.5,
"path": "question_0084/realworld_0014_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 33.5,
"path": "question_0084/realworld_0014_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: John currently shares an apartment with another person, and his bed is the one by the window. These two videos were recorded by him before and after ... | Please select the correct answer from the following options:
A. He did not move anything in either place;
B. He moved things in both places;
C. He only moved things on John's bed;
D. He only moved things in the bathroom.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning p... | B |
question_0085 | (Cross-Video) Memoery Update | [] | Please calculate how many chairs need to be moved to restore the layout of the scene in the second video to match that of the first video. | [
"4",
"3",
"5",
"6"
] | [
[
"question_0085/realworld_0007_0/00:00.00_frame_0.50_0.jpg",
"question_0085/realworld_0007_0/00:00.50_frame_0.50_1.jpg",
"question_0085/realworld_0007_0/00:01.00_frame_0.50_2.jpg",
"question_0085/realworld_0007_0/00:01.50_frame_0.50_3.jpg",
"question_0085/realworld_0007_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 69,
"path": "question_0085/realworld_0007_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 50,
"path": "question_0085/realworld_0007_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Please calculate how many chairs need to be moved to restore the layout of the scene in the second video to match that of the first video. | Please select the correct answer from the following options:
A. 4;
B. 3;
C. 5;
D. 6.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | C |
question_0086 | (Cross-Video) Memoery Update | [] | Amy wore a pair of recording glasses. Two video clips show what Amy saw at two different times (noon and now) in her workplace. She currently has something to throw away, and she wants to know: How many trash bins are there in these rooms, and have these trash bins been moved by the cleaning staff compared to noon? | [
"3 trash bins, two have been moved",
"3 trash bins, all have been moved",
"2 trash bins, none have been moved",
"2 trash bins, one has been moved, one is uncertain",
"2 trash bins, all have been moved",
"3 trash bins, one has been moved"
] | [
[
"question_0086/realworld_0007_1/00:00.00_frame_0.50_0.jpg",
"question_0086/realworld_0007_1/00:00.50_frame_0.50_1.jpg",
"question_0086/realworld_0007_1/00:01.00_frame_0.50_2.jpg",
"question_0086/realworld_0007_1/00:01.50_frame_0.50_3.jpg",
"question_0086/realworld_0007_1/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 50,
"path": "question_0086/realworld_0007_1.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 69,
"path": "question_0086/realworld_0007_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Amy wore a pair of recording glasses. Two video clips show what Amy saw at two different times (noon and now) in her workplace. She currently has som... | Please select the correct answer from the following options:
A. 3 trash bins, two have been moved;
B. 3 trash bins, all have been moved;
C. 2 trash bins, none have been moved;
D. 2 trash bins, one has been moved, one is uncertain;
E. 2 trash bins, all have been moved;
F. 3 trash bins, one has been moved.
. You need t... | A |
question_0087 | (Cross-Video) Memoery Update | [
"question_0087/image_content_1.png"
] | A photographer recorded two video segments of a neighborhood, which may have overlapping content. Focusing on the trees visible in the image <image>, how many of these trees appeared in the frame only during the second recording? | [
"1",
"2",
"4",
"3"
] | [
[
"question_0087/realworld_0018_0/00:00.00_frame_0.50_0.jpg",
"question_0087/realworld_0018_0/00:00.50_frame_0.50_1.jpg",
"question_0087/realworld_0018_0/00:01.00_frame_0.50_2.jpg",
"question_0087/realworld_0018_0/00:01.50_frame_0.50_3.jpg",
"question_0087/realworld_0018_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 29,
"path": "question_0087/realworld_0018_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 28.5,
"path": "question_0087/realworld_0018_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: A photographer recorded two video segments of a neighborhood, which may have overlapping content. Focusing on the trees visible in the image <image>,... | Please select the correct answer from the following options:
A. 1;
B. 2;
C. 4;
D. 3.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0088 | (Cross-Video) Memoery Update | [] | This is a video I took of the room at 9 o'clock and 12 o'clock respectively. How many objects in the bathroom have changed positions? | [
"4",
"2",
"1",
"3"
] | [
[
"question_0088/realworld_0014_1/00:00.00_frame_0.50_0.jpg",
"question_0088/realworld_0014_1/00:00.50_frame_0.50_1.jpg",
"question_0088/realworld_0014_1/00:01.00_frame_0.50_2.jpg",
"question_0088/realworld_0014_1/00:01.50_frame_0.50_3.jpg",
"question_0088/realworld_0014_1/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 33.5,
"path": "question_0088/realworld_0014_1.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 30.5,
"path": "question_0088/realworld_0014_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This is a video I took of the room at 9 o'clock and 12 o'clock respectively. How many objects in the bathroom have changed positions? | Please select the correct answer from the following options:
A. 4;
B. 2;
C. 1;
D. 3.
. You need to choose one letter from A,B,C,D as your answer and provide your detailed reasoning process (based on what information you arrived at this conclusion). Only provide your output(answer and reason) in JSON format: {"answer"... | B |
question_0089 | (Cross-Video) Memoery Update | [] | This is my recording of the same road at two different times in the night with a significant interval. Focusing on the area enclosed by the two power poles, compared to the earlier and later moments, how many originally parked vehicles have left, and how many new vehicles have parked there? | [
"No vehicles have left and no new vehicles have joined",
"1 vehicle left and 1 new vehicle joined",
"1 vehicle left and no new vehicles joined",
"1 vehicle left and 2 new vehicles joined",
"2 vehicles left and 1 new vehicle joined",
"No vehicles left and 1 new vehicle joined"
] | [
[
"question_0089/realworld_0001_0/00:00.00_frame_0.50_0.jpg",
"question_0089/realworld_0001_0/00:00.50_frame_0.50_1.jpg",
"question_0089/realworld_0001_0/00:01.00_frame_0.50_2.jpg",
"question_0089/realworld_0001_0/00:01.50_frame_0.50_3.jpg",
"question_0089/realworld_0001_0/00:02.00_frame_0.50_4.... | [
{
"base_fps": 2,
"end": 61,
"path": "question_0089/realworld_0001_0.mp4",
"start": 0
},
{
"base_fps": 2,
"end": 62,
"path": "question_0089/realworld_0001_1.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task involves comparing two videos of the same scene from different periods, whose timelines are not aligned (the same timecode in the top-left corner is not correspond to the same moment). The central objective is to infer changes in object states from these non-synchronized perspectives. | The first video I provided to you is: <video>. The second video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This is my recording of the same road at two different times in the night with a significant interval. Focusing on the area enclosed by the two power... | Please select the correct answer from the following options:
A. No vehicles have left and no new vehicles have joined;
B. 1 vehicle left and 1 new vehicle joined;
C. 1 vehicle left and no new vehicles joined;
D. 1 vehicle left and 2 new vehicles joined;
E. 2 vehicles left and 1 new vehicle joined;
F. No vehicles left... | A |
question_0090 | Planning | [] | This video shows the layout inside an office. Assume I am now standing somewhere in the room, facing a table corner, with two red chairs on my right and one on my left, all at some distance from me. I want to check the status of the TV and leave the room. What's the fastest way to go? | [
"Turn right and walk to the table corner, then turn left and walk a short distance, the TV is on the right, check it and then keep walking.",
"Turn right and walk a short distance, the TV is on the right, check it and then head back, turn right at the table corner and keep walking.",
"Turn right and walk a shor... | [
[
"question_0090/scannet_fix_0068_0/00:00.00_frame_1.00_0.jpg",
"question_0090/scannet_fix_0068_0/00:01.00_frame_1.00_1.jpg",
"question_0090/scannet_fix_0068_0/00:02.00_frame_1.00_2.jpg",
"question_0090/scannet_fix_0068_0/00:03.00_frame_1.00_3.jpg",
"question_0090/scannet_fix_0068_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 28,
"path": "question_0090/scannet_fix_0068_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video shows the layout inside an office. Assume I am now standing somewhere in the room, facing a table corner, with two red chairs on my right and one on my left, all at some distance from me. I wa... | Please select the correct answer from the following options:
A. Turn right and walk to the table corner, then turn left and walk a short distance, the TV is on the right, check it and then keep walking.;
B. Turn right and walk a short distance, the TV is on the right, check it and then head back, turn right at the ta... | B |
question_0091 | Planning | [
"question_0091/image_content_0.png"
] | <image> Based on the video information, Alex enters the room from the door shown in the picture above, wanting to lie down on the comfortable double sofa to rest and watch TV. How should he go? | [
"After entering the door, go straight ahead, turn right at the end and enter the door of the room, go straight to the sofa after entering",
"After entering the door, turn right then left, then right, walk straight to the end, turn right and go straight to the sofa",
"After entering the door, turn right then lef... | [
[
"question_0091/realestate10k_0016_0/00:00.00_frame_1.50_0.jpg",
"question_0091/realestate10k_0016_0/00:01.50_frame_1.50_1.jpg",
"question_0091/realestate10k_0016_0/00:03.00_frame_1.50_2.jpg",
"question_0091/realestate10k_0016_0/00:04.50_frame_1.50_3.jpg",
"question_0091/realestate10k_0016_0/00... | [
{
"base_fps": 0.6666666666666666,
"end": 175.5,
"path": "question_0091/realestate10k_0016_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image> Based on the video information, Alex enters the room from the door shown in the picture above, wanting to lie down on the comfortable double sofa to rest and watch TV. How should he go? | Please select the correct answer from the following options:
A. After entering the door, go straight ahead, turn right at the end and enter the door of the room, go straight to the sofa after entering;
B. After entering the door, turn right then left, then right, walk straight to the end, turn right and go straight t... | B |
question_0092 | Planning | [] | Amy is a housewife. The activities she does after waking up in the morning are (1) preparing breakfast, (2) taking clothes from the washing machine to dry, (3) turning on the TV, (4) washing up. Please help her plan the sequence to travel the shortest path.(Note: She sleeps in the bed with the pink pattern.) | [
"(4)(3)(1)(2)",
"(4)(1)(2)(3)",
"(1)(2)(3)(4)",
"(3)(2)(4)(1)",
"(4)(3)(2)(1)",
"(3)(4)(1)(2)"
] | [
[
"question_0092/room3d_0230_0/00:00.00_frame_1.00_0.jpg",
"question_0092/room3d_0230_0/00:01.00_frame_1.00_1.jpg",
"question_0092/room3d_0230_0/00:02.00_frame_1.00_2.jpg",
"question_0092/room3d_0230_0/00:03.00_frame_1.00_3.jpg",
"question_0092/room3d_0230_0/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1.000100010001,
"end": 119.98800000000001,
"path": "question_0092/room3d_0230_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Amy is a housewife. The activities she does after waking up in the morning are (1) preparing breakfast, (2) taking clothes from the washing machine to dry, (3) turning on the TV, (4) washing up. Please h... | Please select the correct answer from the following options:
A. (4)(3)(1)(2);
B. (4)(1)(2)(3);
C. (1)(2)(3)(4);
D. (3)(2)(4)(1);
E. (4)(3)(2)(1);
F. (3)(4)(1)(2).
. You need to choose one letter from A,B,C,D,E,F as your answer and provide your detailed reasoning process (based on what information you arrived at this ... | A |
question_0093 | Planning | [] | This video records the layout of my room, and my roommate has a red backpack. Assuming I am now in front of a black suitcase facing the wardrobe, I want to tidy up my quilt before going out. What should I do? | [
"Turn left 90 degrees, walk a distance into the hallway, turn right 90 degrees to tidy my quilt, then turn right 90 degrees and walk straight to leave the room",
"Turn left 90 degrees, walk a distance into the hallway, turn right 90 degrees to tidy my quilt, then turn left 90 degrees and walk straight to leave th... | [
[
"question_0093/scannet_fix_0028_0/00:00.00_frame_1.00_0.jpg",
"question_0093/scannet_fix_0028_0/00:01.00_frame_1.00_1.jpg",
"question_0093/scannet_fix_0028_0/00:02.00_frame_1.00_2.jpg",
"question_0093/scannet_fix_0028_0/00:03.00_frame_1.00_3.jpg",
"question_0093/scannet_fix_0028_0/00:04.00_fra... | [
{
"base_fps": 1,
"end": 58,
"path": "question_0093/scannet_fix_0028_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: This video records the layout of my room, and my roommate has a red backpack. Assuming I am now in front of a black suitcase facing the wardrobe, I want to tidy up my quilt before going out. What should ... | Please select the correct answer from the following options:
A. Turn left 90 degrees, walk a distance into the hallway, turn right 90 degrees to tidy my quilt, then turn right 90 degrees and walk straight to leave the room;
B. Turn left 90 degrees, walk a distance into the hallway, turn right 90 degrees to tidy my qu... | A |
question_0094 | Planning | [] | You want to rest in a cozy yellow-layout bedroom. What is the fastest path from the following three locations: (1) Just came out of the bathroom door (2) Climbed up the stairs from outside (3) Just came out of another bedroom? | [
"Turn right at (1), turn right at (2), turn left at (3)",
"Turn left at (1), turn left at (2), turn right at (3)",
"Turn right at (1), turn right at (2), turn right at (3)",
"Turn left at (1), turn left at (2), turn left at (3)",
"Turn right at (1), turn left at (2), turn right at (3)",
"Turn left at (1),... | [
[
"question_0094/room3d_0047_0/00:00.00_frame_1.00_0.jpg",
"question_0094/room3d_0047_0/00:01.00_frame_1.00_1.jpg",
"question_0094/room3d_0047_0/00:02.00_frame_1.00_2.jpg",
"question_0094/room3d_0047_0/00:03.00_frame_1.00_3.jpg",
"question_0094/room3d_0047_0/00:04.00_frame_1.00_4.jpg",
"ques... | [
{
"base_fps": 1.000100010001,
"end": 199.98000000000002,
"path": "question_0094/room3d_0047_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: You want to rest in a cozy yellow-layout bedroom. What is the fastest path from the following three locations: (1) Just came out of the bathroom door (2) Climbed up the stairs from outside (3) Just came ... | Please select the correct answer from the following options:
A. Turn right at (1), turn right at (2), turn left at (3);
B. Turn left at (1), turn left at (2), turn right at (3);
C. Turn right at (1), turn right at (2), turn right at (3);
D. Turn left at (1), turn left at (2), turn left at (3);
E. Turn right at (1), t... | B |
question_0095 | Planning | [
"question_0095/image_content_1.png"
] | I am wandering in the room shown in the video, and the following image shows the view in front of me. I want to go to the kitchen, how should I go? <image> | [
"Turn around and go straight",
"Turn right and go to the end, then turn right and go straight",
"Turn left and go to the end, then turn left and go straight",
"Insufficient information to make a judgment"
] | [
[
"question_0095/arkitscenes_fix_0042_0/00:00.00_frame_1.00_0.jpg",
"question_0095/arkitscenes_fix_0042_0/00:01.00_frame_1.00_1.jpg",
"question_0095/arkitscenes_fix_0042_0/00:02.00_frame_1.00_2.jpg",
"question_0095/arkitscenes_fix_0042_0/00:03.00_frame_1.00_3.jpg",
"question_0095/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 119,
"path": "question_0095/arkitscenes_fix_0042_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am wandering in the room shown in the video, and the following image shows the view in front of me. I want to go to the kitchen, how should I go? <image> | Please select the correct answer from the following options:
A. Turn around and go straight;
B. Turn right and go to the end, then turn right and go straight;
C. Turn left and go to the end, then turn left and go straight;
D. Insufficient information to make a judgment.
. You need to choose one letter from A,B,C,D as... | C |
question_0096 | Planning | [] | I am currently standing in front of the kitchen sink facing the window, and I want to go to the bedroom to rest. How should I go? | [
"First turn left and walk a distance, then turn left and keep walking",
"First turn right and walk a distance, then turn right and keep walking",
"Turn back and keep walking",
"Insufficient information to determine"
] | [
[
"question_0096/arkitscenes_fix_0042_0/00:00.00_frame_1.00_0.jpg",
"question_0096/arkitscenes_fix_0042_0/00:01.00_frame_1.00_1.jpg",
"question_0096/arkitscenes_fix_0042_0/00:02.00_frame_1.00_2.jpg",
"question_0096/arkitscenes_fix_0042_0/00:03.00_frame_1.00_3.jpg",
"question_0096/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 119,
"path": "question_0096/arkitscenes_fix_0042_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: I am currently standing in front of the kitchen sink facing the window, and I want to go to the bedroom to rest. How should I go? | Please select the correct answer from the following options:
A. First turn left and walk a distance, then turn left and keep walking;
B. First turn right and walk a distance, then turn right and keep walking;
C. Turn back and keep walking;
D. Insufficient information to determine.
. You need to choose one letter from... | D |
question_0097 | Planning | [] | Tony is currently facing the window while washing his hands. If he wants to go upstairs to organize his things, how should he walk to avoid any obstacles? | [
"Turn around and go straight to the front of the table, then turn right and go straight forward to the stairs.",
"Turn right and go straight past the table, then turn right and go straight forward to the stairs.",
"None of the other options are feasible.",
"Turn around and go straight past the table, then tur... | [
[
"question_0097/arkitscenes_fix_0042_0/00:00.00_frame_1.00_0.jpg",
"question_0097/arkitscenes_fix_0042_0/00:01.00_frame_1.00_1.jpg",
"question_0097/arkitscenes_fix_0042_0/00:02.00_frame_1.00_2.jpg",
"question_0097/arkitscenes_fix_0042_0/00:03.00_frame_1.00_3.jpg",
"question_0097/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 119,
"path": "question_0097/arkitscenes_fix_0042_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: Tony is currently facing the window while washing his hands. If he wants to go upstairs to organize his things, how should he walk to avoid any obstacles? | Please select the correct answer from the following options:
A. Turn around and go straight to the front of the table, then turn right and go straight forward to the stairs.;
B. Turn right and go straight past the table, then turn right and go straight forward to the stairs.;
C. None of the other options are feasible... | C |
question_0098 | Planning | [] | As an excellent household robot, you need to execute the owner's instructions as quickly as possible. You have just entered the room, and you need to fetch the toy doll and turn on the TV. What should you do? | [
"Walk forward to the corner to fetch the doll, then walk left to the corner to turn on the TV",
"Walk left to the corner to fetch the doll, then walk right to the corner to turn on the TV",
"Walk forward to the corner to turn on the TV, then walk left to the corner to fetch the doll",
"Walk right to the corne... | [
[
"question_0098/arkitscenes_fix_0036_0/00:00.00_frame_1.00_0.jpg",
"question_0098/arkitscenes_fix_0036_0/00:01.00_frame_1.00_1.jpg",
"question_0098/arkitscenes_fix_0036_0/00:02.00_frame_1.00_2.jpg",
"question_0098/arkitscenes_fix_0036_0/00:03.00_frame_1.00_3.jpg",
"question_0098/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 81,
"path": "question_0098/arkitscenes_fix_0036_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: As an excellent household robot, you need to execute the owner's instructions as quickly as possible. You have just entered the room, and you need to fetch the toy doll and turn on the TV. What should yo... | Please select the correct answer from the following options:
A. Walk forward to the corner to fetch the doll, then walk left to the corner to turn on the TV;
B. Walk left to the corner to fetch the doll, then walk right to the corner to turn on the TV;
C. Walk forward to the corner to turn on the TV, then walk left t... | A |
question_0099 | Planning | [
"question_0099/image_content_0.png"
] | <image>You are currently seated in the position shown in the image and want to retrieve the portrait. Which way should you go? | [
"Turn around and go to the end, then turn right",
"Turn right and go to the end, then turn right",
"Turn around and go to the end, then turn left",
"Turn right and go to the end, then turn left"
] | [
[
"question_0099/arkitscenes_fix_0053_0/00:00.00_frame_1.00_0.jpg",
"question_0099/arkitscenes_fix_0053_0/00:01.00_frame_1.00_1.jpg",
"question_0099/arkitscenes_fix_0053_0/00:02.00_frame_1.00_2.jpg",
"question_0099/arkitscenes_fix_0053_0/00:03.00_frame_1.00_3.jpg",
"question_0099/arkitscenes_fix... | [
{
"base_fps": 1,
"end": 62,
"path": "question_0099/arkitscenes_fix_0053_0.mp4",
"start": 0
}
] | Please answer questions based on video content.
Each video is presented as a frame sequence, with each frame's top-left corner displaying a timestamp ("xx m yy s") that indicates the exact capture time (e.g., 02m20.50s). The timestamps on the initial and final frames correspond to the video's start and end times.
| This task requires you to formulate a plan for a given requirement based on the video information. | The video I provided to you is: <video>. Next, I will ask a question based on the video information. My question is: <image>You are currently seated in the position shown in the image and want to retrieve the portrait. Which way should you go? | Please select the correct answer from the following options:
A. Turn around and go to the end, then turn right;
B. Turn right and go to the end, then turn right;
C. Turn around and go to the end, then turn left;
D. Turn right and go to the end, then turn left.
. You need to choose one letter from A,B,C,D as your answ... | D |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.