In this study, melanoma was detected at an early stage using modified CAT optimization algorithm (MCOA) based on non-convex boundary edge extraction, pixel size, shape and intensity variations on the skin. MCOA can detect skin cancer at an early stage by extracting the non-convex border of the affected region prevent cancer spread. Thus melanoma is curable when detected at an early stage. MCOA extracts image features and obtains non-convex boundaries of melanoma in the skin image. The non-convex boundary region leads to visualization of discriminative features of melanoma based on the region of interest and scaling. The proposed MCOA delineates the affected region through non-convex border extraction and edge detection. An accuracy of 85% was obtained in the detection of melanoma using MCOA, when compared to traditional algorithms.
Keywords
Contour Refinement, Edge Detection, Melanoma, Non-Convex Boundary, Optimization Algorithm.
User
Font Size
Information