Research On Recognition Of Tongues With Toothprint And Relativity Between Sub-Health State And Them

With the development of pattern recognition and digital image processing technology, computer technology has been applied gradually in depth in Traditional Chinese Medicine(TCM). Automation and objectivity of TCM diagnosis can avoid many conventional drawbacks of TCM. As an important aspect of inspection of TCM, tongue diagnosis plays a significant role in TCM diagnoses. The tongue with toothprint, which is an usual pathology mode, has been paid more and more attention by TCM because of its relativity with human sub-health state.This paper mainly studies the recognition and positioning of the tongue with toothprint and the classification of the severe level of the toothprint distribution on tongue in the automated tongue diagnosis system. The relativity between the tongue with toothprint and the human sub-health state, because of whose uncertainty, is further discussed in this paper.First, the recognition algorithm of the tongue with toothprint is studied in this paper. Using the individual methods of the curvature feature and the color feature around the contour of the tongue, the number and location of the toothprint can be obtained. Combining the two methods mentioned above, final recognition results are acquired by means of the comprehensive analysis of the results. Experimental results show that these methods above can obtain good recognition rate, but the misjudgment rate of comprehensive analysis is obviously lower than the other two, which indicates that the combined method is more effective.Next, the classification of the severe level to judge the tongue with toothprint is carried out in this paper. According to the results of toothprint recognition acquired above, features of the tongue with toothprint are extracted by using the Bayes classifier, so that the severe level can be classified. Then satisfactory results are obtained.Finally, using statistical analysis, the relativity between the tongue with toothprint and the human sub-health state is taken. According to the existed data samples of tongue images with the tag of health state, using the extracted features and the severe level of the tongue with toothprint, statistical analysis is carried out by Logistic regression model. The analysis results show that there is significant relativity between the tongue with toothprint and the human sub-health state.