Survival rate of skin cancer is high, if detected early it is curable. So an efficient method is necessary to detect skin lesion at the earliest. The cost of dermatoscope screening for the patient is high, there is a need for an automated system to detect skin lesions captured using a standard digital camera. The automated system is to reduce the percentage of error by choosing the appropriate method in each stage. The features used in the system are extracted byusing GLCM (Gray Level Co-Occurrence Matrix). The output of GLCM is given as the input to SVM (Support Vector Machine) classifier which takes training data, testing data and grouping information which classifies whether given input image is cancerous or non-cancerous. The Cancerous image is taken and the Texture Distinctiveness Lesion Segmentation,Texture means shape or subspace. The pixel variation is used to identify the roughness, smoothness, or bumps or other deformations. The Segmentation is achieved by Morphological Operations and the Sobal filter is used for edge detection technique. The feature extraction is based on ABCD (asymmetry, border, colour and diameter).Classify the Skin Cancer images based on their extracted features. And then types and levels of skin cancer can be classified by TDV (Texture Distinctiveness Value).

Keywords : Segmentation, Classification, TDV, GLCM, Sobal Filter.