Multi Class Skin Diseases Classification Based On Dermoscopic Skin Images Using Deep Learning

Skin diseases are the most common types of health illness faced by people of different age groups. The identification and classification of skin disease problems relies on highly expert doctors and high level instruments which is a time consuming process. To avoid this delay in diagnosis, an automated system is required to identify and classify this skin disease. This paper proposes a convolutional neural network based intelligent system for multi-class skin disease categorization. Three pre-trained deep learning based convolutional neural network models, VGG16, MobileNet, and Inception V3, are used in this study to classify skin disease images. The Dermoscopic image dataset HAM10000 is used for training, validating, and testing. The proposed system is designed, implemented, and tested to classify skin lesion image into one of seven categories. The implementation result of these models using HAM10000 Dataset is obtain as MobileNet accuracy 85.72%,VGG16 accuracy 73.63% and Inception V3 accuracy 75.80%.