A novel fusion approach for segmenting dermoscopy image based on region consistency

Malignant melanoma is among the most rapidly increasing cancers in the world. Image border detection is often the first step to characterize skin lesion for the follow-up computer-aided diagnosis. Existing approaches lack robustness in the face of dermoscopy images varying in size, color, texture, and structure. In this paper, a novel approach is proposed to fuse the segmentation results obtained from different algorithms either in the gray-scale or color space, by discarding the subregions similar to the background skin based on their region consistencies in intensity, size, and texture. The experimental results on the real dermoscopy image set demonstrate that the proposed method can improve the overall performance in terms of both accuracy and robustness.