Digital Tongue Tongue Color Map Analysis Method

1 ObjectiveIn order to investigat the tongue color classification difference at ancient and in modern time. We sorted the tongue color classification in the ancient monographs before Qing danasty. In order to conquer the series of problems in the research on the tongue color analysis at present, eg disagreement of color space, statistics of colorimetric value only on the single parameter, and the limited selection of typical tongue color. We apply the theory of color science and the aproaches of colorimetry to solve above problems and research on the data analysis of tongue color.2 Methods2.1 Theoretical studyThe tongue color classification has been analyzed by the bibliographic methods in the ancient monographs before Qing danasty. Then, using the methods of color science, we analyzed the color three character of the discription of tongue color classification, that is, Ligntness, Chroma, and Hue.2.2 Experimental study2.2.1 Using the data collection method of the tongue color sample manual choosing , we got the different tongue color datum of 706 cases digital picture.2.2.2 Applying the Matlab mathematics software and SAS statistical software, different and complex tongue color datum had been statistical analyzed and displayed in the 3D color space.2.2.3 Using the mathematics algorithm, a single point of centricity had been chosen from several color data by the minimum distance to all the color point in the color space, which had been recognized as the representative tongue color data.2.2.4 Using the mathematics algorithm of polynomial imitation, the space distribution characteristic of different tongue color had been discripted. Then the centrage equations and bouncary equations had been got and been verified.2.2.5 Using the mathematics algorithm, we extracted the typical characteristic color series of different tongue color. Then, we applied the Musell color system and the natural terminology to the discribting the characteristic color.3 Results3.1 Theoretical study3.1.1 The conception of tongue color experienced the development from the color of whole lingual surface to the texture of tongue only. The landmark of distinguishing the tongue coating and the texture of tongue is the Zhi Yi Bi Bian wrote by Li Wenrong in Qing danasty.3.1.2 The tongue color classification catalog appeared by the mark of the Shang Han Guan She Xin Fa wrote by Shen Douyuan, which indicate the diagnosis methods of tongue color completed gradually. 3.1.3 Eight species of tongue color appeared by turns before Qing dynasty. Black tongue and purple tongue both have the description of the texture of tongue and the tongue coating. The appearing of purple tongue was related to the change of green tongue pathogenesis; the deep red tongue was proposed by the school of febrile disease; the blue tongue was appeard often in the serious disease; the white tongue was early proposed by the tongue diagnosis application on the disease caused by the endogenous pathogenic factor.3.1.4 From the color science, the classification and description of the tongue color had been analyzed, and we found that the processing of tongue color recognition began from distinguishing the different color of tongue, that is Hue, gradully to observe the chroma of tongue color, that is Saturation, and Lightness of tongue color. In the thesis, we applied the method of color science firstly on sorting the description terminology of tongue color classification from the hue, saturation and lightness.3.1.5 Comparing with the tongue color classification of Qing dynasty ago, we found that the black tongue and blue tongue had disappered nearly in the modern Chinese medicine, and a few description of tongue color hasn't been used no more, replacing of light dark tongue and dark red tongue.3.2 Experimental study3.2.1 Using the data collection method of the tongue color sample manual choosing , we got the 3D color spatial distribution picture of tongue color compared with the fixed location method of data collection, which coincidence with the tongue color judgement of cilincal doctor.3.2.2 Using the mathematics algorithm, several color data from the same tongue picture has been transformed to a single center spot data. In the color spatial distribution, the choosing points almost overlayed all the area of datum and have afferensed distribution.3.2.3 Comparing the color datum collected from different position,eg the lateral of tongue, the tip of tongue, the tongue coating, the glisten part of tongue surface,the shadow pare of tongue surface, there were the obvious statistic difference and occupied the different space position in the 3D color space.3.2.4 In the CIELAB and the LCH color space, there were the fusiform 3D shape distribution of different tongue color datum, and there also have the satisfied distinguish in the a*-b* plane, the L*-a* plane, the C*-H* plane and the L*-C* plane different tongue color spatial distribution, then3.2.5 We got the center simulating cuve equations and the dividing line equations of different tongue color spatial distribution, then verified their accurate ratio of 81.98～97.88%.3.2.6 Based on the above results, we extracted the typical characteristic color series of different tongue color and made the reference cells of colorchecker of different tongue color. Further, we got the Musell label and the natural terminology of some cells of colorchecker.3.2.7 Comparing the color datum of the amphi-purple tongue and the amphi-dark tongue of the light red tongue and the red tongue and the deep red tongue with the typical light red tongue, the typical red tongue, the light red tongue group and the red tongue group, there were the obvious statistic difference and occupied the different space position in the 3D color space.4 Conclusions4.1 Theoretical study4.1.1 The classification of tongue color experienced the process from simple to mutiply, and then from mutiply to brief. As the application of tongue diagnosis from the Shang-Han disease to the febrial disease, and from the diseases caused by exogenous pathogenic factor to the diseases caused by endogenous pathogenic factor, the classifaction of tongue color became mature and developed to perfection till the later period of Qing dynasty. After our country building up, in order to satisfied the need of TCM teaching, the classification of tongue color had been summarized, which is the process from numerous to brief. However, the brief classification of tongue color can't be satisfied the demand of clinical experience, At present, clinical tongue color discript terminology was applicated confused. Therefore, the terminology of tongue color classifaction and discription need normalization.4.1.2 In the multitude tongue color classification of modern time and before Qing dynasty, the discription of varies of tongue color and amphi-color had the obvious overlap, which induced to the confused discription of tongue color classification and the subjectivity of tongue diagnosis. Then applying the modern color science methods to the quantitative study on the discription of tongue color classification, the normalization of tongue color classification and discription can be benefited.4.2 Experimental study4.2.1 In the quantitative study on the tongue color of digital tongue picture, the disturbed factors of tongue coationg ect should be avoided as far as possible for the tongue color analysis. Compared with the fixed location data collection method and the computer auto-analysing the tongue color, we used the data collection method of the tongue color sample manual choosing, and get the more accuratly analytic results.4.2.2 In order to accuratly caculate the tongue color data, it's need to collect several color data from the different position on the same tongue picture. We proposed a mathematics algorithm of choosing the point of centricity that can be used to transform several color data to a single data.4.2.3 The CIELAB and the LCH color space not only are the uniform distribution vision metric space, but also can ensure the real reappearance and resource sharing for the tongue color datum in the process of transmission and treatment. Therfore, they are the optimal color space for the discription and analysis of tongue color.4.2.4 At frist time, we investigated on the descripation and classification of varies kinds of tongue color in the CIELAB and LCH color space, and got the the center simulating cuve equations and the dividing line equations of different tongue color spatial distribution,which made up the shortage of single parametric statistic analysis of tongue color data at present.4.2.5 At frist time, we get the typical cells colorcheck by extracting the typical characteristic color series of different tongue color based on the distribution of different tongue color in the color space, which made up the dispersed and unsystemic method of extracting the typical tongue color sample directly from the typical tongue picutre.Also, we got the Musell label of some cells of colorchecker, which made the tongue color description of TCM relationship with the international standard.4.2.6 At frist time, we investigated on the distributed characteristic of different typical tongue colors and various kinds of amphi-color tongues in the color space. Also, we proposed that the LCH space transformed from the CIELAB color space was the optimal color space for descripting different hues of amphi-color tongues.