Effective Skin Disease Detection by Analyzing Color and Texture Using Image Processing

Abstract

Diseases related to skin are very contagious and are a worrying problem in the society. It has a great impact on the appearance and mental stability of the patient. Skin diseases should be detected at an initial stage, so that they don’t spread and proper treatment is given for the detected disease. Fungal infection, bacteria, allergy, or viruses are the common causes of skin diseases. The advancement of photonics, lasers, and scientific technologies which are based on medical technology makes detection of skin diseases quickly and with great accuracy also. The only disadvantage it has is that it is very expensive and the patient has to travel all the way to the hospital. Recently, artificial intelligence is taken into consideration for diagnosing skin diseases by using machine learning algorithms and making the use of large amount of data that is available in the health center and hospitals. In this paper, many previous research studies are collected, studied, and reviewed. Within the past ponders, the analysts have analyzed a few frameworks, instruments, and calculations which have been effective in classifying skin infections. Systems have kept faith on methods of image processing and feature extraction that help us predict the skin diseases. To implement skin detection using image processing, we will be extracting features which will help us to classify skin diseases. Due to inappropriate weathers, population, and pollution in some areas, skin diseases are common and spread easily. The work will help detect skin diseases easily and with less time with good accuracy. In this process of detection, initially we take an image of the infected area as an input, and then analyze the image to identify the type of the disease. Our approach does not require any costly or huge instruments but just a camera, computer, or a phone and infected patient’s image. This approach is implemented on feature extraction, the features which we extract are color and texture of the image. We’ve used CNN for image processing and AI for classification. Then multiclass SVM is used; the result is displayed after all the processing. The work is focused on detecting 3 different types of diseases using image processing. The diseases reviewed are Eczema, Psoriasis, and Melanoma.