A Computer-Aided Diagnostic Tool for Melanoma

摘要

Early detection of melanoma, the deadliest form of skin cancer, has the potential to reduce morbidity and mortality. However, clinical diagnosis of melanoma is not trivial even for experienced dermatologists, as it is complex and subjective. Thus, it is necessary to develop an automated computer-aided diagnosis (CAD) system for melanoma, which makes objective judgments based on quantitative measures. In this paper, we present a CAD tool with functionalities of image acquisition, image processing, feature extraction, and diagnosis. The tool provides the capability of detecting melanoma with 92% accuracy, 100% sensitivity, and 92% specificity.