Development of fpga based support vector machine classifier for diagnosis of skin cancer

Melanoma skin cancer cases are observed worldwide more than one hundred and thirty thousand every year. This number can increase dramatically especially in the regions where the white race lives. According to the world health organization, skin cancer can cause the loss of more than sixty-five thousand people every year. Experts underline that early detection and intervention can be life-saving. Therefore, it should be noted that automatic or semi-automatic, machine learning-based hardware architecture that is able to help accelerate the diagnosis of skin cancer would be crucial. In this thesis, it is aimed to develop an independent hardware that can decrease the automated diagnosis complexity of the pathology. Accordingly, FPGA technology was used, and the targeted diagnostic capability was provided by application of machine learning techniques. The results show that the developed platform can diagnose skin lesions that may be harmful with a high hit percentage.