BoneFract: A Bone Fracture Dataset

The dataset allows researchers to use deep learning techniques to detect bone fractures and classify the anatomical region from X-ray images. The underlying research hypothesis is that deep learning models can learn meaningful visual patterns from X-ray images and can be used to accurately distinguish between fracture-positive and fracture-negative cases while identifying the anatomical region affected by a fracture.

The X-ray images in this dataset were individually collected from Peoples University of Medical & Health Sciences for Women (PUMHSW), Nawabshah, Shaheed Benazirabad, Sindh, Pakistan. The images are provided in PNG format and were organized according to anatomical region and fracture status for bone fracture detection and classification research. The dataset consists of 47,931 X-ray images categorized into seven anatomical groups: Arm, Wrist, Hand, Pelvis, Thigh, Lower Leg, Foot.