期刊名称：International Journal of Computer Applications and Information Technology
印刷版ISSN：2278-7720
出版年度：2013
卷号：4
期号：1
页码：9-14
出版社：Mahadev Educational Society
摘要：Skin lesions have consistently had one of the most rapidly increasing incidences of all cancers. Early diagnosis is particularly important. However, even with the help of dermoscopy, differentiating malign and benign lesions is a challenging task. More than that, there are many practical situations where only macroscopic imaging is available. This work focuses on macroscopic images segmentation of black skin lesions. We propose a method combining mathematical morphology and edge detection. After the artifacts extraction stage, we refine the skin lesions by using an opening followed by a dilation algorithm. We then proceed to edge detection. We finally make a comparison of our combination method with existing classical methods.
