Application of PCA and LDA methods on gloss recognition research in TCM complexion inspection

In TCM theory, gloss and moisture are recognized as the Qi; blue, red, yellow, white and black are recognized as the color, the five colors governs blood, the gloss governs sprit and Qi. We may estimate the overabundance or insufficiency of the viscera´s essence by observing the changes of face gloss, which is a very important way to judge the condition and extrapolate the prognosis. However, because of lacking of objective data, the traditional way of observing face gloss mainly depends on clinicians´ subjective appraises, which becomes one of the most important factors that hamper TCM´s development. As a grope and study for the objective TCM, we utilize the computer vision skills and apply the feature extraction methods, like PCA, 2DPCA and (2D)²PCA, LDA to the face samples under 4 color spaces as facial gloss extraction methods. The result indicates that 4 methods have a positive effect on extracting gloss information, LDA may reach to the rate of 88.69% in the HSV color space. This article attempts to explore informatization intellectualization of TCM, and attains positive experimental results. It provides a new method for the TCM observation of gloss quantification examination.