Upload 6 files

HandWritingDigitRecognition.py

@@ -0,0 +1,79 @@
+from PIL import ImageGrab
+import tensorflow as tf
+import numpy as np
+import tkinter as tk
+
+# Load the models that you have already trained
+model = tf.keras.models.load_model("./model/model.h5")
+model.load_weights("./model/model_weights.h5")
+
+# Create the tkinter window
+root = tk.Tk()
+root.title("Handwritten Digit Recognition")
+
+# Create the main canvas with black background color
+canvas = tk.Canvas(root, width=280, height=250, bg='black')
+canvas.pack()
+
+screen = tk.Label(root, text="Draw a number", font=("Helvetica", 24))
+screen.pack()
+
+
+# Function to handle drawing on the canvas
+def start_draw(event):
+    global last_x, last_y
+    last_x, last_y = event.x, event.y
+
+
+def draw(event):
+    global last_x, last_y
+    x, y = event.x, event.y
+    canvas.create_line((last_x, last_y, x, y), fill="white", width=10)
+    last_x, last_y = x, y
+
+
+# Function to predict the drawn digit
+def predict_digit():
+    x = root.winfo_rootx() + canvas.winfo_x()
+    y = root.winfo_rooty() + canvas.winfo_y()
+    x1 = x + canvas.winfo_width()
+    y1 = y + canvas.winfo_height()
+
+    # There is a buggy area here
+    # When I try to capture the canvas image sometimes it captures wrong places.
+    # So I tried to fix it manually but, it still doesn't work perfect:(
+    img = ImageGrab.grab((x+31, y+38, x1, y1))
+
+    # See the captured image
+    # img.show()
+
+    img = img.convert('L')
+    img = img.resize((28, 28))
+    img_array = np.array(img)
+    img_array = img_array.reshape(1, 28, 28, 1) / 255.0  # Normalize input
+    prediction = model.predict(img_array)
+
+    # Hold the best prediction
+    predicted_digit = np.argmax(prediction)
+    screen.config(text="Predicted digit: " + str(predicted_digit))
+
+
+# Function to clear the canvas
+def clear_canvas():
+    canvas.delete("all")
+    screen.config(text="Draw a number")
+
+
+# Bind mouse events
+canvas.bind("<Button-1>", start_draw)
+canvas.bind("<B1-Motion>", draw)
+
+# Predict button
+predict_button = tk.Button(root, text="Predict", command=predict_digit)
+predict_button.pack()
+
+reset_button = tk.Button(root, text="Clear", command=clear_canvas)
+reset_button.pack()
+
+# Start the main loop
+root.mainloop()

model training.ipynb

@@ -0,0 +1,183 @@
+{
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
+    "colab": {
+      "provenance": []
+    },
+    "kernelspec": {
+      "name": "python3",
+      "display_name": "Python 3"
+    },
+    "language_info": {
+      "name": "python"
+    }
+  },
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": 2,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "1vuQoR42AF_N",
+        "outputId": "5c473da6-0bf9-42ac-8e7a-993d4522623f"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
+            "11490434/11490434 [==============================] - 0s 0us/step\n",
+            "Epoch 1/10\n",
+            "468/468 [==============================] - 100s 210ms/step - loss: 0.3633 - accuracy: 0.8851 - val_loss: 0.0526 - val_accuracy: 0.9829\n",
+            "Epoch 2/10\n",
+            "468/468 [==============================] - 96s 204ms/step - loss: 0.1228 - accuracy: 0.9621 - val_loss: 0.0363 - val_accuracy: 0.9879\n",
+            "Epoch 3/10\n",
+            "468/468 [==============================] - 98s 209ms/step - loss: 0.0898 - accuracy: 0.9730 - val_loss: 0.0298 - val_accuracy: 0.9896\n",
+            "Epoch 4/10\n",
+            "468/468 [==============================] - 99s 210ms/step - loss: 0.0719 - accuracy: 0.9786 - val_loss: 0.0198 - val_accuracy: 0.9935\n",
+            "Epoch 5/10\n",
+            "468/468 [==============================] - 100s 213ms/step - loss: 0.0680 - accuracy: 0.9791 - val_loss: 0.0193 - val_accuracy: 0.9938\n",
+            "Epoch 6/10\n",
+            "468/468 [==============================] - 97s 207ms/step - loss: 0.0590 - accuracy: 0.9816 - val_loss: 0.0175 - val_accuracy: 0.9942\n",
+            "Epoch 7/10\n",
+            "468/468 [==============================] - 101s 216ms/step - loss: 0.0547 - accuracy: 0.9830 - val_loss: 0.0181 - val_accuracy: 0.9941\n",
+            "Epoch 8/10\n",
+            "468/468 [==============================] - 101s 215ms/step - loss: 0.0511 - accuracy: 0.9845 - val_loss: 0.0155 - val_accuracy: 0.9946\n",
+            "Epoch 9/10\n",
+            "468/468 [==============================] - 100s 214ms/step - loss: 0.0476 - accuracy: 0.9858 - val_loss: 0.0188 - val_accuracy: 0.9937\n",
+            "Epoch 10/10\n",
+            "468/468 [==============================] - 100s 213ms/step - loss: 0.0444 - accuracy: 0.9858 - val_loss: 0.0182 - val_accuracy: 0.9948\n"
+          ]
+        },
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "<keras.src.callbacks.History at 0x7ef01c2c8520>"
+            ]
+          },
+          "metadata": {},
+          "execution_count": 2
+        }
+      ],
+      "source": [
+        "import tensorflow as tf\n",
+        "import numpy as np\n",
+        "\n",
+        "# Load the pre prepared data from TensofFlow\n",
+        "(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()\n",
+        "\n",
+        "# Preprocess the data\n",
+        "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
+        "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
+        "x_train = x_train.astype('float32')\n",
+        "x_test = x_test.astype('float32')\n",
+        "x_train /= 255\n",
+        "x_test /= 255\n",
+        "\n",
+        "# Define a CNN model with data augmentation\n",
+        "model = tf.keras.models.Sequential([\n",
+        "    tf.keras.layers.Conv2D(64, (3, 3), activation=\"relu\", input_shape=(28, 28, 1)),\n",
+        "    tf.keras.layers.MaxPooling2D(2, 2),\n",
+        "    tf.keras.layers.Conv2D(64, (3, 3), activation=\"relu\"),\n",
+        "    tf.keras.layers.MaxPooling2D(2, 2),\n",
+        "    tf.keras.layers.Flatten(),\n",
+        "    tf.keras.layers.Dense(512, activation=\"relu\"),\n",
+        "    # Dropout to avoid from the overfitting\n",
+        "    tf.keras.layers.Dropout(0.5),\n",
+        "\n",
+        "    # There are 10 different (0-9) output.\n",
+        "    # So use 10 units dense layer with softmax activation\n",
+        "    tf.keras.layers.Dense(10, activation=\"softmax\"),\n",
+        "])\n",
+        "\n",
+        "# Compile the model\n",
+        "model.compile(optimizer='adam',\n",
+        "              loss='sparse_categorical_crossentropy',\n",
+        "              metrics=['accuracy'])\n",
+        "\n",
+        "# Data Augmentation\n",
+        "datagen = tf.keras.preprocessing.image.ImageDataGenerator(\n",
+        "    rotation_range=10,\n",
+        "    width_shift_range=0.1,\n",
+        "    height_shift_range=0.1,\n",
+        "    shear_range=0.1,\n",
+        "    zoom_range=0.1\n",
+        ")\n",
+        "\n",
+        "datagen.fit(x_train)\n",
+        "\n",
+        "# Train the model with data augmentation\n",
+        "model.fit(datagen.flow(x_train, y_train, batch_size=128),\n",
+        "          steps_per_epoch=len(x_train) / 128,\n",
+        "          epochs=10,\n",
+        "          validation_data=(x_test, y_test))\n"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "loss, acc = model.evaluate(x=x_test, y=y_test)\n",
+        "print(\"Loss:\", loss)\n",
+        "print(\"Accuracy:\", acc)"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "3O5b1ShcIE7-",
+        "outputId": "6ef12f6b-eaa2-4c85-97da-db2ee7d81b6e"
+      },
+      "execution_count": 3,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "313/313 [==============================] - 4s 13ms/step - loss: 0.0182 - accuracy: 0.9948\n",
+            "Loss: 0.018189573660492897\n",
+            "Accuracy: 0.9947999715805054\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "# Save the model and model weights to a avaliable path\n"
+      ],
+      "metadata": {
+        "id": "2pNqtozRN5wj"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "model.save(\"/content/model/model.h5\")\n",
+        "model.save_weights(\"/content/model/model_weights.h5\")"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "7Fp3_wYEKMF-",
+        "outputId": "14213539-b8bf-4774-8916-5a0ea28b23a2"
+      },
+      "execution_count": 4,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stderr",
+          "text": [
+            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
+            "  saving_api.save_model(\n"
+          ]
+        }
+      ]
+    }
+  ]
+}

model.py

@@ -0,0 +1,51 @@
+import tensorflow as tf
+import numpy as np
+
+# Load the pre prepared data from TensofFlow
+(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()
+
+# Preprocess the data
+x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)
+x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)
+x_train = x_train.astype('float32')
+x_test = x_test.astype('float32')
+x_train /= 255
+x_test /= 255
+
+# Define a CNN model with data augmentation
+model = tf.keras.models.Sequential([
+    tf.keras.layers.Conv2D(64, (3, 3), activation="relu", input_shape=(28, 28, 1)),
+    tf.keras.layers.MaxPooling2D(2, 2),
+    tf.keras.layers.Conv2D(64, (3, 3), activation="relu"),
+    tf.keras.layers.MaxPooling2D(2, 2),
+    tf.keras.layers.Flatten(),
+    tf.keras.layers.Dense(512, activation="relu"),
+    # Dropout to avoid from the overfitting
+    tf.keras.layers.Dropout(0.5),
+
+    # There are 10 different (0-9) output. 
+    # So use 10 units dense layer with softmax activation 
+    tf.keras.layers.Dense(10, activation="softmax"),  
+])
+
+# Compile the model
+model.compile(optimizer='adam',
+              loss='sparse_categorical_crossentropy',
+              metrics=['accuracy'])
+
+# Data Augmentation
+datagen = tf.keras.preprocessing.image.ImageDataGenerator(
+    rotation_range=10,
+    width_shift_range=0.1,
+    height_shift_range=0.1,
+    shear_range=0.1,
+    zoom_range=0.1
+)
+
+datagen.fit(x_train)
+
+# Train the model with data augmentation
+model.fit(datagen.flow(x_train, y_train, batch_size=128),
+          steps_per_epoch=len(x_train) / 128,
+          epochs=10,
+          validation_data=(x_test, y_test))

model/model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ae84744a8bc0e7eabf0dd277f89c23dc561c937deb12b13ac8cd0c13a33fed8
+size 10393504

model/model_weights.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:240bd369c26bf4659c7d53de65e8e1f2177b369843187521a795999527ff1587
+size 3472256

requirements.txt

@@ -0,0 +1,4 @@
+PIL
+tensorflow
+numpytkinter
+numpy