Bone Tumor X-ray Images with Annotated Localizations

Annotated bone tumor X-ray dataset for segmentation and localization using deep learning. The dataset contains medical images with bounding box or segmentation mask annotations. It was sourced from Kaggle, but the author, organization, and specific volume are unknown.

Use Cases

• Train object detection models to localize bone tumors using the provided bounding box annotations.
• Develop segmentation models to precisely outline tumor boundaries from the pixel-level mask annotations.
• Benchmark the performance of different deep learning architectures on medical image localization tasks.
• Create educational tools for visualizing tumor characteristics in X-ray images.

Strengths

• Contains annotations specifically for object detection and segmentation tasks.
• Focuses on a specific medical domain (bone tumors) for targeted model development.

Limitations

• The total number of images and annotations is unknown, making scale assessment impossible.
• Potential class imbalance or bias is unverified due to lack of sample data and demographics.
• The annotation quality and consistency are not described or validated.

Provenance

Source

Kaggle

Collection Method

Annotated X-ray images collected for deep learning tasks; specific gathering method is unknown.