Cancer is one of the dreadful diseases in today’s world. Skin cancer is one among those. Manual methods have been available, still a contribution towards computer aided diagnostic in the field of medicine plays a challenging role. The objective of the paper is to detect and classify the skin image and find if it is a melanoma or non melanoma images. This paper includes the method for image enhancement as adaptive median filter, segmentation as a hybrid method and the feature extraction to be a hybrid method and finally the KNN classifier is used to classify the images. The total of 600 images is taken and the efficiency is said to be 96.8%. © IJSTR 2020.