Advanced earlier melanoma detection algorithm using colour correlogram

Melanoma is a most dangerous form of skin cancer that develops from the pigment producing cells known as melanocytes. Melanoma skin cancers are also known as malignant melanoma. Recent studies show that the death rates of melanoma patients depend on the various stages of cancer, so early detection and treatment of melanoma implicate higher chances of cure. Now most of the existing skin lesion analysis system use ABCDE parameters for feature extraction. But these methods have lot of drawbacks. In this paper an advance earlier melanoma detection algorithm is proposed using colour correlogram and texture analysis. Bayesian classifier is used to detect the abnormal skin cells with colour correlogram and SFTA feature vectors. The system is successfully tested with the dermoscopic dataset and the experimental results show that the combination of colour correlogram and texture analysis give better results with an accuracy of 91.5%.