Developing automatic diagnostic tools for the early detection of skin cancerlesions in dermoscopic images can help to reduce melanoma-induced mortal-ity. Image segmentation is a key step in the automated skin lesion diagnosispipeline. In this paper, a fast and fully-automatic algorithm for skin lesionsegmentation in dermoscopic images is presented. Delaunay Triangulation isused to extract a binary mask of the lesion region, without the need of anytraining stage. A quantitative experimental evaluation has been conductedon a publicly available database, by taking into account six well-known state-of-the-art segmentation methods for comparison. The results of the experi-mental analysis demonstrate that the proposed approach is highly accuratewhen dealing with benign lesions, while the segmentation accuracy signi-cantly decreases when melanoma images are processed. This behavior led usto consider geometrical and color features extracted from the binary masksgenerated by our algorithm for classication, achieving promising results formelanoma detection.
Skin lesion image segmentation using Delaunay Triangulation for melanoma detection
BLOISI, Domenico Daniele;
2016-01-01
Abstract
Developing automatic diagnostic tools for the early detection of skin cancerlesions in dermoscopic images can help to reduce melanoma-induced mortal-ity. Image segmentation is a key step in the automated skin lesion diagnosispipeline. In this paper, a fast and fully-automatic algorithm for skin lesionsegmentation in dermoscopic images is presented. Delaunay Triangulation isused to extract a binary mask of the lesion region, without the need of anytraining stage. A quantitative experimental evaluation has been conductedon a publicly available database, by taking into account six well-known state-of-the-art segmentation methods for comparison. The results of the experi-mental analysis demonstrate that the proposed approach is highly accuratewhen dealing with benign lesions, while the segmentation accuracy signi-cantly decreases when melanoma images are processed. This behavior led usto consider geometrical and color features extracted from the binary masksgenerated by our algorithm for classication, achieving promising results formelanoma detection.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.