Automatic skin lesion segmentation via iterative stochastic region merging

TitleAutomatic skin lesion segmentation via iterative stochastic region merging
Publication TypeJournal Article
Year of Publication2011
AuthorsWong, A., J. Scharcanski, and P. Fieguth
JournalIEEE Transactions on Information Technology in Biomedicine
Pagination929 - 936
Keywordsiterative, lesion, region merging, skin cancer, stochastic

An automatic method for segmenting skin lesions in conventional macroscopic images is presented. The images are acquired with conventional cameras, without the use of a dermoscope. Automatic segmentation of skin lesions from macroscopic images is a very challenging problem due to factors such as illumination variations, irregular structural and color variations, the presence of hair, as well as the occurrence of multiple unhealthy skin regions. To address these factors, a novel iterative stochastic region-merging approach is employed to segment the regions corresponding to skin lesions from the macroscopic images, where stochastic region merging is initialized first on a pixel level, and subsequently on a region level until convergence. A region merging likelihood function based on the regional statistics is introduced to determine the merger of regions in a stochastic manner. Experimental results show that the proposed system achieves overall segmentation error of under 10% for skin lesions in macroscopic images, which is lower than that achieved by existing methods.