A Suitable Approach for Classifying Skin Disease Using Deep Convolutional Neural Network

Abstract

Skin cancer is a form of cancer that develops in skin tissue and can harm nearby tissues, resulting in disability, and even result in death. Skin cancer’s detrimental consequences can be reduced and controlled with an accurate diagnosis and prompt, effective treatment. Using a CNN (Convolutional Neural Network), a system is constructed in this study that is capable of automatically distinguishing between skin cancer lesions. HAM10000 image dataset has been deployed for conducting the task of categorizing the skin lesions for seven classes; MNV, MLN, BKT, DFM, BCC, ACTK, and VSL. The quantity of photos has also increased by applying various techniques such as image augmentation. The model has shown satisfactory results and the final accuracy achieved was 90.34% and the validation accuracy of 90.87%. The loss was nominal and the model is capable of classifying the skin lesions in the correct category. This model can be used in the medical area along with the research area for the classification of other skin issues.