Heel Dataset

Heel Bone X-Ray Dataset consists of 3,956 X-ray images of the foot, primarily focused on detecting and classifying heel bone diseases. The images were obtained from Kirkuk General Hospital in Digital Imaging and Communications in Medicine (DICOM) format and converted to JPG format using the MicroDicom tool.

The dataset is labeled into three distinct classes:

Normal: 1,842 images Heel Spur: 1,316 images Severe (Heel Spur Complications): 798 images

All labels were assigned by a specialist in orthopedics and traumatology and were cross-verified by a radiology specialist for validation, ensuring high-quality annotations. This dataset is valuable for training models on medical image classification, specifically targeting foot-related conditions.

Dataset Size: 3,956 images

Image Format: JPG (originally in DICOM format)

Source: Kirkuk General Hospital

https://www.kaggle.com/datasets/osamahtaher/heel-dataset/data

Paper: HeCapsNet: An enhanced capsule network for automated heel disease diagnosis using lateral foot X-Ray images.

Taher, Osamah, and Kasım Özacar. "HeCapsNet: An enhanced capsule network for automated heel disease diagnosis using lateral foot X‐Ray images." International Journal of Imaging Systems and Technology 34.3 (2024): e23084.