Skin Disease Classification using Image Preprocessing and Machine Learning

Efficient and timely identification of skin disease is important in medical diagnosis. This work aims to implement the possibility of skin disease detection using image preprocessing and machine learning. Image resizing and enhancement are implemented in image preprocessing. In enhancement, images are modified using brightness, contrast, and sharpness adjustments. Here, some bottom and top pixels values are saturated. Decision tree and Support Vector Machine based algorithms are considered to implement the machine learning. Feature selection plays a vital role in machine learning. In this work, features are selected from 3 color models and histogram count. Color models considered in this work are RGB, HSI, and LAB. Three important features are selected from each color model. Performance metrics used in this work are accuracy, precision, and recall. Experimental results indicate that the proposed method gives better accuracy.