Analysis Of The Features Of Tongue Coating And Tongue Texture In Children With Food Accumulation Disorder

In recent years,Internet and medical has been widely concerned by all walks of life,and the medical diagnosis method based on artificial intelligence has been gradually recognized by the public.Tongue diagnosis,as an indispensable routine examination in TCM,plays an important role in clinical diagnosis of traditional Chinese medicine.Its unique diagnosis mode reflects the traditional experience and characteristics of TCM diagnosis.The ability of self-control of young children is poor,and the uncontrolled diet is easy to cause "accumulated food".Children with accumulated food will generally suffer from anorexia,anorexia,bad breath,discomfort in intestines and stomach,restless sleep,fever,sweating and cold hands and feet,and even fever in severe cases.Therefore,it is very important to detect and accurately judge the symptoms of accumulated food in time.In view of the popularity of smart phones in China and the care of parents for children,parents hope that they can check whether children suffer from accumulated food disease through smart phones at any time and anywhere,and take timely measures to eliminate the adverse effects of the disease on children.This paper puts forward a set of solutions for the diagnosis of childrens bulimia by mobile terminal(1)Aiming at the problem that there is no good way to deal with the color correction of mobile terminal image,this paper proposes an evaluation method of lighting conditions,that is,to evaluate the current lighting conditions by measuring the color vector of tongue image under the same lighting conditions with or without flash lamp,classifies the current lighting conditions by support vector machine method,and then classifies the current lighting conditions through multiple color correction methods In this paper,the polynomial model of traditional color correction is deeply studied.Through the design and comparison of several polynomial model experiments,a better polynomial model is determined,which lays a foundation for the subsequent tongue image processing.(2)Aiming at the problems of poor segmentation accuracy and easy to be interfered by irrelevant regions,the traditional algorithm is easy to be interfered by irrelevant regions.Two steps of tongue segmentation based on deep learning are designed: using yolov3 algorithm to recognize the tongue,that is to obtain ROI(region of)of tongue The output of yolov3 algorithm is used as the input of u-net algorithm,and the tongue can be segmented from the original image accurately.Compared with the traditional segmentation algorithm,the experimental results show that the final segmentation effect is far better than the traditional tongue segmentation algorithm,which is the pathological characteristics of bulimia the judgment of sign eliminates the interference of external environment.(3)In view of the problems such as the great harm of childrens bulimia,low vigilance of parents and complicated medical procedures,this chapter uses the lab color space to make qualitative and quantitative analysis on the projection of childrens tongue images of bulimia and normal tongue images in lab space,and determines the distribution of tongue images with mild,severe and normal conditions in lab color space,and how to The experimental results were compared with the results of tongue images labeled by old Chinese medicine,and good results were obtained.