Human Bone Fracture C17 Dataset

This dataset is a curated collection of medical images of human bone fractures, categorized into 17 distinct fracture types. It has been compiled to support research in medical image classification, deep learning and clinical decision support systems. Fracture classification is an important aspect of orthopedic diagnostics, as different fracture types require different treatment approaches. By providing a diverse, well-labeled dataset, this collection can be used to train and benchmark machine learning models for automatic fracture detection and classification.

The dataset contains 17 classes of bone fractures:
1. Avulsion fracture
2. Closed (simple) fracture
3. Comminuted fracture
4. Compression (crush) fracture
5. Fracture dislocation
6. Greenstick fracture
7. Hairline fracture
8. Impacted fracture
9. Intra-articular fracture
10. Longitudinal fracture
11. Oblique fracture
12. Open (compound) fracture
13. Pathological fracture
14. Segmental fracture
15. Spiral fracture
16. Stress fracture
17. Transverse fracture

Image Type: X-ray, radiology
Format: JPG / PNG / WEBP
Organization: Each fracture type is stored in its own folder
Intended Use: Research, medical AI development, and educational purposes

The dataset is built using publicly available medical and radiology images from:
Kaggle: Bone Break Classifier Dataset (https://www.kaggle.com/datasets/amohankumar/bone-break-classifier-dataset)
Radiopaedia: Stress Fracture Article (https://radiopaedia.org/articles/stress-fracture-2)
Shutterstock: Medical and radiology image database (https://www.shutterstock.com)