Colour Image Segmentation using Optimal Fuzzy Clustering

J.L. Baker, D. Campbell, and A. Bodnarova (Australia)


Colour Segmentation, Optimal Fuzzy Clustering, FuzzySystems, Computer Vision


This paper presents a novel approach to pixel based colour image segmentation. The method presented utilises the hue, intensity and saturation representation, from which a hue histogram is constructed. Analysis is performed on the hue histogram using an optimising fuzzy clustering tech nique to partition the histogram. The resulting histogram partitions are used to segment the hue plane, achieving an image level segmentation. This paper demonstrates the effectiveness of the optimal fuzzy clustering approach to colour image segmentation by accurately locating regions of homogenous colour whilst minimising the rate of over segmentation.

Important Links:

Go Back