Research Of Skin Pigmentation Detection Technology Based On Image Analysis

The pigmentation characteristic parameters are important indicators inmeasuring skin lesions’ degree. The feature extraction of skin pigmentation byimage analysis technique, it has many advantages, such as not contact with theskin, fast, reliable and the measurement data information can be reused. In thefield of medical pathology diagnosis has caused wide public concern. In thispaper, using the image processing technology such as image enhancement, imagesegmentation, to extract the characteristics of the pigmentation, specific studiesare as follows:First of all, through analyzing the characteristics of the human skin tissuestructure and skin images found the pigmentation image contrast difference isvery big.On this basis, using the gray mean square error method to classifysplash image, So that we can further identified as "fuzzy" pigmentation imageenhancement processing. It can’t solve the problem of local high brightness ofskin image effectively because of some problems in one of three ways by theanalysis of gray level transformation, median filtering, and histogramequalization enhancement experiment. So it combine the wavelet transform withthe homomorphic filtering to enhance the image contrast effectively, and solvethe problem of partial skin image brighter.Secondly, through study has found that the Snake model for using the greedyalgorithm can maintain stability in the process of contour extraction and haslower computational complexity. Therefore, in this paper, we put forward usingthe Snake model, realize the splash image contour extraction. Adopting themethod of GVF Snake model correction to solve the problem of the contourcurve does not converge to deep concave boundary, by to improve the quality ofsegmentation result. Finally, using the8-chain code method extracts the skin pigmentation areaboundary and other geometric parameters. In order to reduce thecalculating error area of the skin pigmentation, we took a half pixel unitprocessing area boundary. So constructed with boundary characteristics,geometric features, moment invariant features, color spots parameters, thusforming the objective evaluation index system, by comparing the change of thecorresponding indexes before and after treatment, the preliminary evaluation oftherapeutic is given effect.