Skin Cancer Classification Framework Using Enhanced Super Resolution Generative Adversarial Network and Custom Convolutional Neural Network

Melanin skin lesions are most commonly spotted as small patches on the skin. It is nothing but overgrowth caused by melanocyte cells. Skin melanoma due to abnormal surge of melanocytes. The number patients suffering from cancer observably rising globally. Timely and precise identification crucial for lowering mortality rates. An expert dermatologist required handle cases using dermoscopy images. Improper diagnosis can cause fatality patient if it not detected accurately. Some classes come under category benign while rest malignant, causing severe issues diagnosed at an early stage. To overcome these issues, Computer-Aided Design (CAD) systems proposed which help reduce burden giving them accurate There several deep learning techniques that implemented classification. In this experimental study, we have a custom Convolution Neural Network (CNN) Human-against-Machine (HAM10000) database publicly accessible through Kaggle website. designed CNN model classifies seven different present in HAM10000 database. achieves accuracy metric 98.77%, 98.36%, 98.89% protocol-I, protocol-II, protocol-III, respectively, Results our models also assimilated with literature were found be superior than them. enhance performance metrics, initially pre-processed Enhanced Super Resolution Generative Adversarial (ESRGAN) gives better image resolution images smaller size.