Human ringworm detection using wavelet energy signature

We propose an application based experimental work to identify the ringworm images from a set of human skin images. Our approach deals with the 3-level decomposition of the skin images by the Daubechies (DB), Coiflet (CF), Biorthogonal (BO) and Discrete Meyer (DM) wavelets and extraction of the corresponding energy signatures. The discriminatory energy signatures from the different wavelet decomposed approximation and detail subbands at each level of resolution are used to tabulate the training and testing databases. The binary classifier, Support Vector Machine (SVM) is then deployed to detect the ringworm images successfully.