Dynamic 3D reconstruction of tongue surface based on photometric stereo technique

Abstract

Tongue inspection is an important part of the diagnostic methods in traditional Chinese medicine. In the process of the objective research for tongue diagnosis, the tongue information we obtained from the analysis of two dimensional (2D) tongue image is limited. The aim of this study is to reconstruct dynamic three-dimensional (3D) surface of tongue based on photometric stereo technique. We utilized a highlight excluding algorithm in order to reduce the effects of the highlight areas on the reconstruction results. The overall outline and detailed 3D surface of the tongue could be reconstructed in this way with an average relative error of 7.24%. The motion simulation of tongue was achieved via 3D morphing for better diagnosis. Our experience with this work demonstrated that the reconstruction method for the tongue was accurate and reliable for both clinic and research purposes. Further works still need to be done to improve the accuracy of the reconstruction result.