Support Vector Machine Based Melanoma Skin Cancer Detection

In Present Scenario, one of the life threatening disease which causes human death is Skin cancer. The cause of skin cancer is due to the abnormal growth in melanocytic cells. Due to genetic factors and exposure of ultraviolet radiation , Melanoma appears on the skin as brown or black in colour. Early diagnosis can cure this skin cancer completely. The traditional method to detect the skin cancer is Biopsy which is invasive and painful. This method of laboratory testing consumes more time. To resolve the above issues, diagnosis of skin cancer is developed based on computer aided. The proposed system uses four phases to detect the skin cancer. First, it uses Dermoscopy to capture the skin image. Next step is to pre-process the image . After the step of pre-processing, it is segmented which is followed by feature extraction with unique features from the segmented lesion. A last, these features were given to a supervised classifier named support vector machine (SVM) to classify whether the given image is as normal image or melanoma diseased skin image . The experimental result shows SVM classifier offers more accuracy compared to existing technique.