Published Paper Details:

This project addresses the demand for an intelligent and rapid classification system of skin diseases - Acne, Melanoma, Psoriasis, Rosacea, Vitiligo using highly-efficient convolutional neural network. Diseases like Melanoma can be prove to be fatal if not identified early. Thus, identifying the diseases at early stages is critical to save lives. Towards this goal we propose a system that uses recent deep Convolutional Neural Network learning methods that are capable of classification of skin disease. The model is trained with labelled data of the skin diseases. Using feature extraction, the features of the diseases like the diameter, color and border are identified. The features are then summarized while retaining the important information. Then, the image of the skin lesion is loaded into the model along with various factors like the patient's history, drinking or smoking habits and results of the tissue testing. These factors help provide additional information about the patient to produce more accurate results. The system then classifies the type of skin disease along with the severity of the diseases based on the trained data. Automated systems capable of detecting diseases could save lives, costs and reduce needless interference.