Title :
Colour segmentation for the analysis of pigmented skin lesions
Author :
Schmid, Ph ; Fischer, S.
Author_Institution :
Signal Process. Lab., Swiss Federal Inst. of Technol., Lausanne, Switzerland
Abstract :
A colour segmentation scheme based on two-dimensional histogram analysis and the fuzzy c-means clustering technique is presented. The method is applied to digitised epiluminescence microscopy images of pigmented skin lesions and is used for the extraction of all homogeneously coloured regions. Methods to reduce the influence of small structures, hairs, bubbles, light reflection, and other artefacts are proposed as pre- and post-processing steps
Keywords :
image colour analysis; bubbles; colour segmentation; digitised epiluminescence microscopy images; epiluminescence microscopy slides; filtering; fuzzy c-means clustering technique; hairs; homogeneously coloured regions; light reflection; pigmented skin lesions; postprocessing; preprocessing; small structures; two-dimensional histogram analysis;
Conference_Titel :
Image Processing and Its Applications, 1997., Sixth International Conference on
Conference_Location :
Dublin
Print_ISBN :
0-85296-692-X
DOI :
10.1049/cp:19970983
