Research On Automatic Identification System Of Tcm Constitution Based On Facial Image Feature

TCM holds that the human body is an organic whole, which needs to carry on health care and disease prevention according to individual constitution. So, the changes of facial morphology, look and body constitution are linked to a certain extent. And the current study mainly focused on the relationship between looking and viscera disorder or other specific conditions. In this thesis, we take the face image of the subject, and then to use the digital image processing and pattern recognition technology to initially complete the automatic analysis of face consultation. And the research of this thesis improves the accuracy of physical identification, provides more theoretical basis for physical identification, and has important practical significance for the TCM objective development and TCM Constitution identification.Generally speaking, this paper mainly completed the following aspects of the work:Firstly, the human face image preprocessing is completed. In the process of image acquisition, the human face image is not uniform because of the human body posture or the angle of the shot. In order to eliminate the interference of other unstable factors, and reduce the error, it is necessary to carry out the image preprocessing. In this paper, the specific location of human face in the image is obtained by using the method of skin color detection, face location and so on.Secondly, the color features and texture features related to the physical types of traditional Chinese Medicine are extracted. The position of the forehead, nose and cheeks of the face is located, then the color feature and texture feature are extracted, and the feature of human face is transformed into the vector of digital form. In the color feature extraction, the image is converted to HSV color space according to the nonlinear transformation formula, and then three components, H, S and V, of the facial sub image are extracted as the color feature. For the texture feature, the texture feature is extracted by using the gray level co-occurrence matrix and wavelet transform, which makes the texture feature better anti disturbance ability.Thirdly, an objective study of relationship between facial image features and traditional Chinese Medicine Constitution was carried out. On the basis of the above research, the support vector machine is used to classify the color feature and texture feature, and to realize the identification of TCM Constitution. Respectively, using grid times calendar and particle swarm optimization method to optimize the kernel parameter and penalty factor C, and combining with the cross validation method to get the optimal parameters. And then, analyze the internal relations between TCM surface image and TCM Constitution.Finally, TCM constitution automatic identification systems based on surface characteristics is developed. Building a whole life cycle health management platform, and designing face automatic analysis module, we have completed the automatic identification of TCM constitution, and provided a new basis for the comprehensive and healthy physical examination of traditional Chinese Medicine.