Aiming at the problems in the color space information

the global distribution and the local regions and edges information of an image are considered separately appearing in traditional color image segmentation

a hierarchical color image segmentation algorithm based on Fuzzy ART model is proposed. The algorithm frame mainly involves the effective utilization of the luminance distribution

the edge information and the color space information. A set of hierarchical features are extracted from the image

and making use of the ability of stable and fast incremental learning and memorizing of Fuzzy ART model

the image is partitioned into several regions hierarchically and the hierarchical representation of the image is formed. According to the hierarchical representation of the image

good segmentation results can be obtained. Compared with the FFCM algorithm

better results can also be obtained.