Research And Implementation On Tongue Image Segmentation And Color Correction Technology

Tongue diagnosis is the important component of traditional Chinese medicine,but there are shortcomings,just rely on subjective observation of the doctor to determine the condition,lack of objective basis,hinder the sustainable development of traditional Chinese medicine.Through modern information technology,a quantitative analysis of tongue,can effectively avoid subjectivity,so objective tongue diagnosis in traditional Chinese medicine is the inevitable trend of the development of Chinese medicine today.This paper is based on the objectification research of the tongue image of Chinese medicine,focuses on the segmentation and color correction of the tongue image,and the research tasks are as follows:In order to realize the segmentation and color correction of the tongue image,this paper studies a variety of image pretreatment methods,including darkening,noise reduction,histogram equalization and sharpening.Aiming at the bright spot problem,the anti-spot of the tongue surface is detected by the fixed threshold method,and the anti-light points are removed by using the image darkening.The Gaussian filter and bilateral filtering algorithm are studied for the possible noise in the acquisition process,which can protect the edge of the tongue while removing the noise;In order to effectively detect the edge of the tongue,the contrast between the tongue and the background is enhanced by means of histogram equalization.In order to improve the quality of the tongue,protruding the details of the tongue,the Laplace sharpen method is studied,and the sharpening process of the tongue image is realized.For efficient and stable on tongue segmentation,this paper studies the watershed algorithm,graph cut algorithm and edge detection algorithms,through simulation experiment,comparing the segmentation results,stability and maneuverability of these algorithms,choose the edge detection algorithm as the method of segmentation;aiming at the color distortion of the tongue image,after finishing the color correction of the acquisition equipment and display equipment,a variety of color correction methods are studied,and the color reduction experiment is carried out.Finally,the partial least squares method is used to solve the problem of tongue color bias.In the end,the software development is carried out in the application layer involved in the image pretreatment,tongue image segmentation and color correction.Using C++ design and implement a software system capable of completing the functions of tongue acquisition,pretreatment,segmentation and color correction,and each function is tested,the diagnosis interface is called to complete the diagnosis of the tongue image at last.