Dermoscopic Image Analysis for Skin Cancer Detection Using Deep Features and Support Vector Machines

With more than 5,000,000 newly diagnosed cases reported in the US alone each year, skin cancer has grown to be a significant public health concern. Despite the fact that the mortality rate is high, the survival rate is above 95% when identified early. This is the motivation behind our work to develop a method for early skin cancer diagnosis that could save the lives of people. The project’ s goal is to develop a skin cancer detection model that can distinguish between different types of skin cancer and help with early cancer detection. This study focuses on skin cancer detection using dermoscopic images, utilizing the HAM10000 dataset. It employs two deep learning models: MobileNet V2 and EfficientNet B0. Subsequently, deep features are extracted from both models. By combining the deep features from both models, a remarkable accuracy of 98.7<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> is achieved.