Font Size: a A A

Features Extraction Of Tongues' Color With Its Application On Nephropathy And The Research In Tongue Shape Classification

Posted on:2007-11-12Degree:MasterType:Thesis
Country:ChinaCandidate:J S WuFull Text:PDF
GTID:2178360212966981Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Traditional Chinese Medicine Tongue Diagnosis (TCMTD) obtains information through observing the tongues' characteristics of patients to get their state of health, and then diagnoses the state of illness. It is one of the most valuable and widely used diagnostic methods in Traditional Chinese Medicine (TCM).The research about TCMTD objectivization and quantification enables computer to assist doctor in the process of TCM diagnosis, and it wants to be carried forward around the world further. A lot of diagnoses errors will be brought on if the sampled tongue image is different from true tongue. Therefore, the quality of sample gathering is very important. It is necessary to gather tongue samples under good environment, reducing errors as few as possibly. In this thesis, we first introduce and analyze the knowledge of color spaces, clustering methods and the Analytic Hierarchy Process (AHP) model. Then, through lots of experiments, we choose the best collecting environment which is the steadiest one in the three kinds of the feasible environments. We gather the nephritic tongue samples in this environment. All above has played an important role in the next research.This thesis is the first tentative research in nephritic tongue recognition. We present a novel color clustering method based on pixels on the wholly tongue surface. Using different typical tongue samples which selected by TCM experts, tongue color ranges can be gotten through clustering. We propose an improved K-Nearest Neighbors (KNN) method for matching every pixel of nephritic tongue samples. Afterwards, we can get color ratio of tongue as color features for recognition. This process accords with TCM diagnosis process and people's subjective judgments on tongue color.This thesis is also the first tentative research on tongue shape. We get tongue marginal points by Snake algorithm. We present a novel algorithm named the minimal width tongue shape correction for tongue shape correction. The Analytic Hierarchy Process which is widely used in Operational Research is first introduced in tongue shape classifying in this thesis. We made some improvements on AHP algorithm and applied it in tongue shape classification.
Keywords/Search Tags:tongue image, color characteristic, C-means, tongue shape, Analytic Hierarchy Process
PDF Full Text Request
Related items