RECOGNITION OF PSORIASIS FEATURES VIA DAUBECHIES D8 WAVELET TECHNIQUE

摘要

This paper presents a study in an efficient methodology for analysis and characterization ofdigital images psoriasis lesions using Daubechies D8 wavelet technique. The methodology is based onthe transformation of 2D Discrete Wavelet Transform (DWT) algorithm for Daubechies D8 at first levelto obtain the coefficients of the approximations and details sub-images. For classification method,statistical approach analysis is applied to identify significance difference between each groups ofpsoriasis in terms of mean and standard deviation parameter. Results performances are concluded byobserving the error plots with 95% confidence interval and applied independent T-test. The testoutcomes have shown that approximate mean and standard deviation parameter can be used todistinctively classify erythroderma from the other groups in consistent with visual observations of the error plots. Whilst, in order to discriminate guttate from the other groups, standard deviationparameters for horizontal, vertical and diagonal can be utilized. Based on the results, plaque isdistinguishable with guttate and erythroderma by using standard deviation vertical sub-imagesparameter. Results of Daubechies D8 is compared with study done previously by using Daubechies D4and Daubechies D12 in order to observe the reliability of the results in Daubechies families. Theresultant parameters can be used to design computer-aided system in diagnosis the skin lesion ofpsoriasis.