Statistical tool to analyse psoriasis disease

Abstract

Computer Aided Diagnosis (CAD) is a rapidly developing newlineframework that provides solutions to many complex problems in the field of newlineMedical Image Processing. Diseases such as dermatological problems have newlinewide range of factors and affects human beings irrespective of their age and newlinegender which has major impact on the quality of their life. It is more helpful newlinefor diagnosis and it supports dermatologists in their treatment procedure. The newlinemain focus of this study is on psoriasis diseases and its types. newlinePsoriasis is a clinically heterogeneous lifelong skin disease. It affects newline60 million people globally. It affects both males and females under the age of newline40 years. It affects the scalp, face, intertriginous areas, genitals, hands, feet, and newlinenails. newlineIn the first study, a framework is developed to classify six types of newlinepsoriasis using Multiclass Particle Swarm Optimization (MPSO). Initially newlinepreprocessing operation is done and segmentatation is used by Color and newlineTexture prior based segmentation. Next 31 texture, 38 color and 14 shape newlinefeature descriptors are extracted. The promising 11 feature vectors are selected newlineusing Genetic algorithm and given input to the classifier with an accuracy of newline95.7% newlineIn the second study, a model is designed for diagnosing psoriasis and newlineits type using Convolutional Neural Network (CNN). Initially, Color and newlineTexture feature based segmentation approach is used to segment psoriasis newlineaffected regions. Image augmentation is done to extend the data size by using newlinerandom flip, rotation between 0and 65, zooming by 2%, shearing by 9% newline