Comparison of Machine Learning Algorithms for Skin Disease Classification Using Color and Texture Features

Abstract

Machine learning algorithms are being used widely in biomedical fields for segmentation and diagnosis. These algorithms use features derived from images as input to make a decision. So, choosing proper feature extraction methods combined with suitable machine learning (ML) algorithms is very important to achieve good classification accuracy. During the literature survey, we found that there is a lack of information about machine learning algorithms for skin disease classification. To address this problem, we have collected Chronic Eczema, Lichen planus and Plaque psoriasis images using a digital camera and extracted Red, Green and Blue (RGB) color features and Gray Level Co-occurrence Matrix (GLCM) texture features. Different combinations of features with four popular ML algorithms were considered to compare classifier performances. Out of the four algorithms tested, Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM) showed highest classification accuracy.