A Comparative Analysis of Skin Cancer Detection based on SVM, ANN and Naive Bayes Classifier

Skin cancer cases have grown worldwide, and it is by far the most common of all cancers. It is caused by the extended exposure to harmful radiations from the sun. Skin cancer can be broadly classified into two categories as melanocytic and non-melanocytic (benign). Most basic way to detect the threat of skin is by visual investigation performed by a specialist dermatologist employing a set of specific clinical tools. This paper describes novel image processing approaches for skin cancer detection using dermoscopy images. The motive of this work is to investigate and prefer an algorithm for skin cancer determination that can group the lesions as malignant or benign melanoma in terms of accuracy, sensitivity and specificity. Initially, the image samples showing melanoma are obtained by using dermoscope and segmented. It is then given for feature extraction and the results are classified using Support Vector Machine (SVM), Artificial Neural Network (ANN) and Naive Bayes Classifiers.