Medical Image Analysis For Detection And Prediction Of Skin Diseases Using CNN

Abstract

Skin disease is one of the most unpredictable and common diseases in our country. It is caused by bacteria, allergies, viruses, fungal infections, etc. Since skin diseases are common and the patients are in large numbers, medical care is required and it should not be ignored. It is very difficult to detect and diagnose due to its complexity. With the advancement in technology, it is conceivable to identify the type and intensity of skin diseases accurately. In dermatology, for finding the condition of the skin, expensive investigations have to be carried out for the diagnosis of the disease. This paper proposes a technique that utilizes analysis of images that extracts the maximum information required for diagnosis by utilizing appropriate data. They are the position or location of disease and segmentations at deformities, or by utilizing the assessable features of appearance indicators from image. A decision system that utilizes manifold classifiers similar to the feature-based neural networks technique is used. Based on the classifier output weight the final decision system is designed. Classification can be improved for better diagnosis based on calculated accuracy for better decisions.