Tongue Image Segmentation Using Deep Learning for Diagnosis in Traditional Chinese Medicine

Tongue image segmentation is important in tongue diagnosis, oral disease diagnosis, and health monitoring in traditional Chinese medicine (TCM). Traditional image segmentation methods require a large amount of data with limitations. Deep learning methods, especially the UNet algorithm architecture remarkably perform in tongue image segmentation. We employed the UNet-based tongue image segmentation method and explored its application using tongue image datasets. The performance in TCM tongue diagnosis was verified in experiments.