Research on Separation Method of Tongue Body and Coating Based on Double Light Sources

Xuemin Wang,Yuanting L,Ruiyun Wang,Xiaozuo Lu,Peng Zhou
DOI: https://doi.org/10.13494/j.npe.20150098
2016-01-01
Abstract:Separation of tongue body and coating is the key to realizing automatic diagnosis in tongue in-spection. However, tongue body and coating are complex, and most existing separation methods, which are based on difference in color, cannot applicable in various situations. For example, staggered thick and thin coating or yellow and white coating cannot be well separated with these methods. Regarding this problem, this paper designs and develops a system to separate tongue body and coating based on double light sources. The influence of white light and pure green light on the separation of tongue body and coat-ing is investigated. Considering the color features of tongue images under white light or pure green light, the tongue images are segmented by Snakes algorithm with different initial contour extraction methods, in-cluding automatic algorithm and artificial interaction algorithm. Mutual information theory for image regis-tration and clustering method for classification are adopted to separate tongue body and coating. Results show that the separation result of pure green light image is preferable to that of the white light image, pro-viding new reference for tongue inspection method in traditional Chinese medicine.
What problem does this paper attempt to address?