Image Detection of Rare Orthopedic Diseases Based on Explainable AI

Qi-Xiang Zhang, Shun-Ping Wang, Yu-Wei Chan, Chih-Hung Chang; Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, 2024, pp. 654-659

Abstract


Image detection has significant application value in medicine, especially in detecting Muller-Weiss Disease (MWD) in orthopedic X-ray images. Traditional manual interpretation methods can be influenced by subjective factors and individual experience, and they can be time-consuming and labor-intensive. In this study, by utilizing advanced object detection models like YOLOv8, we can automatically and accurately identify specific structures and abnormalities in the images, providing real-time feedback, significantly improving physicians' diagnostic accuracy. Furthermore, the use of the Grad-CAM technique to generate heatmaps enhances the interpretability of the model's decisions, helping physicians understand the basis for the model's judgments, further boosting confidence and accuracy in diagnosis. Therefore, image detection plays a critical role in medical image diagnosis, potentially improving diagnostic efficiency and enhancing healthcare quality.

Related Material


[pdf]
[bibtex]
@InProceedings{Zhang_2024_WACV, author = {Zhang, Qi-Xiang and Wang, Shun-Ping and Chan, Yu-Wei and Chang, Chih-Hung}, title = {Image Detection of Rare Orthopedic Diseases Based on Explainable AI}, booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops}, month = {January}, year = {2024}, pages = {654-659} }