An Efficient Detection Framework for Linear Skin Lesions with Pigmentary Disorders

Linear skin lesions have several anatomical and causative factors, which are associated with numerous pigmentary disorders. The automatic detection of such linear lesions in clinical images improves the diagnosis process to distinguish between the potential complications. In this paper, we have proposed an efficient detection framework for linear skin lesions. This framework is based on the linear and curvilinear patterns of the cutaneous pigmentary signs related to these lesions. The proposed detection framework can detect either single or multiple lesion patterns. The detection process is achieved by discovering the line segments of the lesions' objects and finding the layout of their configurations. Several examples of pigmentary disorders of linear diseases are presented in order to evaluate the detection framework, such as Linea Nigra and Striae.