Melanoma segmentation using bio-medical image analysis for smarter mobile healthcare

Abstract

Dermoscopy is an excellent method of detecting melanoma in its early stages. Skin is the principal organ of human body. It covers bones, muscles and all parts of the body. Melanoma is rare, but it is the most dangerous form of skin cancer. It is curable if it is detectedin its early stages. Digital dermoscopy help dermatologists in theexamination of cancerous skin lesions. It enables doctors to capture microscopic images of moles by using amobile phone, corresponding application or any handy scope device. Segmentation is used to divide the image into different segments. Segmentation, classification and feature extraction are the three fundamental stages of arecognition system that helps in matching the analysis of skin lesion. Melanoma occurs due to the presence of Melanocytes in the body. With the use of dermoscopy, the dermatologists can examine individual lesions more closely. In our paper, we have proposed an approach that can automatically preprocess the image and then segment the lesion.The gradient magnitude of the image is calculatedto filter the images. We have marked the foreground objects to segment the lesion from background precisely. The proposed technique is tested on European dataset of dermoscopic images. Results of segmented images are compared with other competitors to exhibit the superiority of the recommended approach. Matlab 2016a is used for successful simulation of the experiment.