Geometric-based nail segmentation for clinical measurements

Abstract A robust segmentation method that can be used to perform measurements on toenails is presented. The proposed as the first step in a clinical trial objectively quantify incidence of particular pathology. For such an assessment, it necessary distinguish nail, which locally appears similar skin. Many algorithms have been used, each leverages different aspects toenail appearance. We Hough transform locate tip toe and estimate nail location size. Subsequently, we classified super-pixels image based their geometric photometric information. Thereafter, watershed delineated border nail. was validated using 348-image medical dataset, achieving accuracy 0.993 F-measure 0.925. considerably across samples, with respect factors shape, skin pigmentation, illumination conditions, appearance large regions affected by condition.