EDGE-BASED IMAGE SEGMENTATION FOR SKIN LESIONS ON ARMS
Authors: Chang Jia Woei and Lau Hui Keng
Image segmentation is the process of partitioning a digital image into multiple regions or sets of pixel. It is an important basic process in image analysis. Skin lesions is defined as a superficial growth or patch of the skin that does not resemble the area surrounding it. Image segmentation for skin lesions aided machine in reading and recognising area of skin lesions in an effective way. Edge-based detection is used for detecting and recognizing skin lesions from an image using the proper step of image processing which start from image pre-processing, segmentation and recognition of skin lesions. Edges is the boundaries of object surfaces in a scene often lead to oriented localized changes in intensity of an image. Example of edge-based segmentation methods are Sobel operator, Prewitt operator, Robert’s operator and Canny Edge Detection. However, different method applied may influence the feature extraction and the final lesion classification. For this work, the focus is in edge-based image segmentation to identify regions of interest, to study which method in edge-based image segmentation is the most accurate. Four edge-based image segmentation techniques will be implement in Matlab and the accuracy of the edge-based segmentation techniques of skin lesions on arms is compared based on the Normalized Probabilistic Rand (NRP) index. To increase the accuracy of the result, some pre-processing or filtering method will be added before segmentation process to remove artefacts on the original images such as hairs and noise that can influence the detection of skin lesion.