Skin lesion segmentation using Gray Level Co-occurance Matrix

© 2016 IEEE. Skin lesions screening is an effective method for early detection of melanoma. Mostly, melanoma appears as hyper-pigmented area relative to the surrounding skin. Lesion segmentation is an indispensable step for skin lesions analysis. Automated segmentation is used to assist the dermatologist to isolate the suspicious lesion from the surrounding background. Iterative Otsu's method is state-of-art segmentation technique and it has acceptable accuracy. However, iterative methods suffer same drawback, they are time consuming and no guarantee for convergence to the best solution before the maximum iterations limit reached. This paper presents a novel segmentation algorithm using GLCM (Gray Level Co-occurrence Matrix). Segmentation masks extracted by proposed method are compared to human-expert extracted ground truth. The proposed method consists of three major stages, preprocessing, segmentation, and post processing. The proposed method achieved a specificity rate of 98.62%, precision of 96.25% and sensitivity of 80.8%.