Abstract

This study presents an efficient way to segment the skin cancer images. A novel method is proposed that combines colour and texture for the segmentation of skin lesions from unaffected skin region in an image. The distributions of colour and texture features provide a platform for the discrimination of skin lesions. The segmentation results are evaluated quantitatively by means of a comparative experiment on a set of skin cancer images. The evaluation of the proposed method is based on the comparison with Live Wire segmentation technique. The results indicate that the proposed methodology proved effective and efficient for the skin cancer image segmentation.