README.md

---
license: apache-2.0
task_categories:
- image-classification
language:
- en
---

# Pectus Excavatum Dataset

## Dataset Description

This dataset contains images of pectus excavatum cases with various demographic information and risk levels. The dataset is intended for medical research and classification purposes.

### Labels

The dataset includes the following label categories:
- Age: adult, child
- Ethnicity: asian, white, hispanic, undetermined
- Gender: male, female
- Risk Level: high-risk, low-risk

### Data Splits

- creation_date: 2025-01-22T14:12:49.587991 images
- original_size: 741 images
- split_stats: N/A images
- class_weights: N/A images
- excluded: N/A images

## Dataset Structure

The dataset is organized into three splits:
- `train/`: Training set
  - `images/`: Training images
  - `labels.csv`: Training labels
- `val/`: Validation set
  - `images/`: Validation images
  - `labels.csv`: Validation labels
- `test/`: Test set
  - `images/`: Test images
  - `labels.csv`: Test labels

Additional files:
- `split_metadata.json`: Contains metadata about the dataset splits
- `test_set_validation_report.json`: Validation report for the test set

### Labels Format

Each `labels.csv` file contains the following columns:
- image_id: Unique identifier for each image
- original_path: Original path of the image
- new_path: New path in the processed dataset
- original_category: Original category label
- age: Age category
- ethnicity: Ethnicity category
- gender: Gender category
- risk_level: Risk level category

## Citation

If you use this dataset in your research, please cite it appropriately.

## Acknowledgements

Thanks to all contributors and medical professionals who helped in creating and labeling this dataset.