A comprehensive survey on image-based computer aided diagnosis systems for skin cancer

摘要

Malignant melanoma is the deadliest form of skin cancer. In 2013 around 14,509 melanoma cases were found in the United Kingdom and the rate is increasing ever since. Melanoma can be easily treatable if detected in early stages. Clinical as well as automated methods are being used for melanoma diagnosis. Image-based computer aided diagnosis systems have great potential for early malignant melanoma detection. In this paper we review state of the art in computer aided diagnosis system and examine recent practices in different steps of these systems. Statistics and results from the most important and recent implementations are analyzed and reported. We compared the performance of recent work based on different parameters like accuracy, dataset, computational time, color space, machine learning technique etc. and summarized them in table format for better understanding of emergent researchers in the field of computer aided skin diagnosis systems. Research challenges regarding the different parts of computer aided skin cancer diagnosis systems are also highlighted.