Abstract:

Psoriasis is a persistent skin disease which inflames the skin producing red and thickened areas with silvery scales. The aim of this work is to do a segmentation of the main structures of the skin images. Our algorithm first isolates a set of zones that certainly correspond to lesion plaques based on chromatic information, and then expands these zones to achieve an accurate segmentation of plaques through a Bayesian modeling. Gabor filters are still able to detect the texture features. The Bayesian rule is used for distinguishing between skin pixels and non-skin pixels The performance of the Gabor filter for non-periodic patterns is tested in apple quality inspection and face recognition Keeping this in mind, we use Gabor filters in this work to differentiate non-periodic scaling patterns from normal skin patterns


Index Terms—Psoriasis, Bayesian Classifiers, Gabour Fliters, Scaling Patterns Segmentation of skin Images;