Mathematical morphology aided shape, texture and colour feature extraction from skin lesion for identification of malignant melanoma

Among the skin cancers, malignant melanoma is of by far the most deadly type. For the early diagnosis of melanoma, computer-aided diagnosis systems for the classification of dermoscopic images have been developed. For accurate image classification, the segmentation of skin lesion area is of prime importance. In this work, mathematical morphology has been used for segmentation for proper identification of skin lesion area. Different shape, texture and color features have been extracted from a set of dermoscopic images and malignant melanomas have been classified using SVM classifier with 85.71% sensitivity.