1. The method is simple
2. The data set is difficult to obtain (not open source), it is convenient to publish papers, it is estimated that it is
If you need a data set, you can contact =. =
References
Tongue image segmentation algorithm based on DCNN and fully connected CRF