Research And Application Of Tongue Coating And Body Separation

Traditional Chinese medicine tongue diagnosis is to observe and analyze the patient’s tongue image from a certain angle under the appropriate environmental conditions,combined with the relevant theory of traditional Chinese medicine and the doctor’s own clinical experience to diagnose the patient’s condition.Traditional Chinese medicine tongue diagnosis is easily affected by the objective environment and the doctor’s own experience,and it is time-consuming and laborious,and its diagnosis results can not be effectively preserved for a long time.Therefore,it is very important to introduce objective indicators into traditional Chinese medicine tongue diagnosis treatment plan.The objectification of tongue diagnosis of traditional Chinese medicine is based on the research of traditional Chinese medicine theory and the introduction of artificial intelligence technology.Its main content is to let computer and other equipment automatically recognize and analyze the input tongue image,and output the diagnosis results.The objectification of TCM tongue diagnosis mainly includes three steps: image segmentation of tongue body,separation of tongue coating and body,and design of automatic tongue diagnosis system.This paper focuses on the above three aspects and makes the following research:Tongue image segmentation.The original tongue image usually contains not only the complete tongue image,but also the image of cheek,lip and tooth,so it is necessary to remove the redundant part.In this paper,grabcut algorithm is used to realize tongue image segmentation,which requires a small amount of user interaction to achieve segmentation.The experimental results verify the applicability of the algorithm for tongue image segmentation,and lay the foundation for the next step of tongue coating and body separation.Tongue coating and body separation.Tongue coating and body separation is to divide the tongue image into two parts: tongue coating image and tongue body image.An improved k-means algorithm for tongue coating and body separation model TCBS is proposed.Firstly,considering RGB,HSV and Lab three common color space models,two components of the three color space models are selected by scoring to form a dual channel tongue image,and the pixel value of the dual channel tongue image is taken as theclustering sample points of the model;then,the initial clustering center point is determined by the peak value of the distribution histogram of the single channel tongue image;finally,the initial clustering center points are used to cluster to realize the separation of tongue coating and body.The experimental results show that TCBS model is better than other methods.Design of automatic tongue diagnosis system.A small automatic tongue diagnosis system based on Python Django framework is designed.The system can automatically upload tongue images,segment tongue body images and automatically separate tongue coating and body.Combined with relevant tongue diagnosis images in database,the possible diseases of patients can be preliminarily obtained and diagnosis suggestions are given,which can provide reference for the diagnosis of TCM doctors.