A Rudimentary Plaque Lesion Identification Using Combination of Color Models

This paper discusses the effort of discriminating plaque psoriasis skin lesions using Artificial Neural Network (ANN) for dermatological early diagnosis based on color representations. For any digital acquired images, colors can be identified numerically, for example, with respect to the unique RGB, HSV and YCbCr pixel indices. Previous work have produced intelligent identification models for selected psoriasis lesion such as plaque, using Artificial Neural Network (ANN) based on each of these individual color models only. However in this work, an identification model using combination of various colors (CoVC) is proposed. The input parameters for training an ANN classifier with supervised Levenberg-Marquardt (LM) algorithm are made up of combination of various color models. Outcomes of this work have revealed that the performance of optimized CoVC model with respect to sensitivity, specificity and accuracy has outclassed previous intelligent identification models when validated at a threshold of ±0.5.