Analysis of Skin disease techniques using Smart Phone and Digital Camera Identification of Skin Disease

Abstract

Skin diseases are a serious health issue that affects a large number of individuals. In recent years, with the fast advancement of technology and the use of various data mining approaches, dermatological predictive classification has become increasingly predictive and accurate. It is more help to dermatologist to identify the disease, As a result, the development of machine learning approaches capable of efficiently. The purpose of this study is making an application of identification skin disease images by using the machines learning method, Support Vector Machine (SVM), and KNN techniques. The image processes and machine learning is performed early detection of skin diseases. The aim of this study is determined the classification of skin diseases in humans. Each skin disease has symptoms. It has five skin diseases such as Acne, Psoriasis, Wrath, Psoriasis, and Ulcer. We have collected 314 skin disease images from the government of hospital, Aurangabad with the help of mobile camera and Sony HD camera. Gaussian Filter is used for image pre-processing. The segmentation method is used for K-Means Clustering and the feature extraction method are used for feature extraction. We have used Haar feature, color feature, FCM, OS-FCM, GLCM and LBF features for classifications. Based on the result, the SVM is given 92% accuracy for haar feature, FCM and OS-FCM. and KNN classifier, K-Means are given 89% and 89% accuracy using mobile phone camera dataset. The SVM, KNN and K-Means are given 91%, 87% and 89% accuracy respectively using Sony HD camera dataset. SVM is given good result in both dataset.