Objective Study Of Human Tongue Based On The Mobile Image

The current Internet health care has been identified as the next major growth point of the national economy.In this thesis,the mobile Internet and traditional Chinese medicine are combined,the collection of mobile phone tongue to carry out segmentation,color correction and attribute determination research.The main research work is as follows:1)Thw tongue image collected by mobile devices,often in a very complex light conditions.In this thesis,the algorithm of LBP and Adaboost is used to realize near real-time tongue detection.At the same time,the deep learning algorithm based on the residual network is used to obtain the segmentation mask of the tongue image and the probability of the image including the tongue image.The algorithm runs on the GPU server contains the Nvidia GeForce GTX 1080,and the running speed is real time.2)In contrast to the color correction algorithm of tongue color,this paper adopts the idea of color correction algorithm based on Retinex theory,combined with the gray world algorithm.In the experimental results and visual conditions have reached a better color correction effect.In this thesis,the tongue color correction algorithm can better remove the shadow,retain the original tongue color.3)This article lists the results of some object-oriented research in practice.The tongue is sorted by the thickness of the tongue and the degree of purple.While exploring the definition of yang index associated with human physique.Finally,based on the previous numerical definition,this thesis gives a more detailed relationship between tongue image data and human health.For users who are in sub-health,this thesis also gives a combination of Chinese medicine formula.