Study On Analysis Method Of Digital Tongue Picture And Its Application On Tooth-Marked Tongue

Tongue diagnosis, as an important diagnostic method in Traditional ChineseMedicine(TCM), needs to be researched for objectivity and digitalization like otherdiagnostic methods. However, tongue diagnosis may be a breakthrough for diagnosticresearch on objectivity and digitalization because it is more objective than other diagnositicmethods. With the development of technology of computer and image analysis, manyresearches have been made and some equipments or softwares have been worked out, butthere are many problems that need to be solved urgently in research on digitalization andobjectivity of tongue diagnosis such as how to understand and evaluate these new values inclinic of TCM. In this research, the distributing law of digital characters of tongue picturesand some commonsense problems such as color model and description of tongue colorhave been explored, and these methods and technology of digitalization of tongue picturehave been applicated on tooth-marked tongue to explore the values and means in clinic ofTCM. The equipment and software of analysis on digital tongue picture in this research isprovided by Lab of Analysis on Signal and Information, Beijing Industrail Universitybecause its equipment and software is more advanced according to pertinent literatures. And this research includes mainly two clinical investigations. (1) The study on distributing law and methods of digital characters of tonguepicture. General checkup people is studied in first investigation in collective ways includingtongue picture results defined by clincal doctors, digital values of tongue picturescomputered by software and symptoms questionaire filled by check-up people. As a result,884 samples have been studied by statistics with two research focuses. Firstly, differentcolor models have been compared and its meaning in clinic has been explored. Secondly,the law and descriping method of distributing status of tongue color have been studied. 1. The study on different color model in digitalization of tongue picture. Methods:RGB and Lab values of tongue color of 884 check-up people have been computed bysoftware of tongue image, and the values of YUV, YCrCb and HSI color model have beenobtained through transformation in color space. Then the values of color models abovehave been compared with tongue color diagnosed by clinic doctors. Results: The values ofL, b and S are correlated significantly to tongue color by clinic doctors. There is asignificant difference in value L and a between light red tongue and red tongue, asignificant difference of value L and b between light red tongue and dark red tongue, asignificant difference of value a between red tongue and dark red tongue. Conclusion: Labmodel is an appropriate color model in digital tongue diagnosis of Traditional ChineseMedicine. 2. The study on distributing law of digital characters of tongue pictures. The digitalcharacters of tongue picture such as the values of L, a, b of tongue texture color and tonguecoating color, area and thickness of tongue coating, crack index and humidity in middletongue have been studied and compared between different sex, age and tongue pictureresults judged by clinical doctor in second part. According to the results, the value a intongue texture color of male check-up people is higher signifcantly than the female whilefemale L and b are higher than male. Meanwhile, the crack index in male is higher thanfemale. And the results above show that the male have more chance of excessive syndromeand heat syndrome than the female. In change of age, the value b of tongue texture colordecreases with age while the value a, area and thickness of tongue coating increaserespondenlly, which shows there are some changes such as blood stasis and excessive heatwith age, and the function of the spleen and stomach and water metabolism in body becomeworse when people are aged. Comparison to the color judged by clinical doctors, it showsthat values L, a and b have some differences between two tongue colors. Moreover, 3 Dscantterplot is recommonded to describe clearl...