Automatic segmentation of skin cancer images using adaptive color clustering
Ilea, Dana E. and Whelan, Paul F. (2006) Automatic segmentation of skin cancer images using adaptive color clustering. In: CIICT 2006 - China-Ireland International Conference on Information and Communications Technologies, 18-19 October 2006, Hangzhou, China.
Full text available as:
This paper presents the development of an adaptive image segmentation algorithm designed for the identification of the skin cancer and pigmented lesions in dermoscopy images. The key component of the developed algorithm is the Adaptive Spatial K-Means (A-SKM) clustering technique that is applied to extract the color features from skin cancer images. Adaptive-SKM is a novel technique that includes the primary features that describe the color smoothness and texture complexity in the process of pixel assignment. The A-SKM has been included in the development of a flexible color-texture image segmentation scheme and the experimental data indicates that the developed algorithm is able to produce accurate segmentation when applied to a large number of skin cancer (melanoma) images.
Archive Staff Only: edit this record