Based On The Analysis Of Fuzzy Clustering Color Image Segmentation

After a deep research and analysis of the actuality and problems of image segmentation based on fuzzy clustering, an approach called as EFCM(Edge-based ). This method use simple, and the sensitive edge operator, carries on the marginal check to the image, obtains in the image all edge information; uses the improvement the FCM algorithm to take the division foundation, completes the image the thick division; takes the fuzzy clustering by the thick division's region number the category number, the thick division various regions center to take the initial cluster center, proposed that makes the fuzzy clustering to the non-edge part, reduces the cluster iteration the number of times, also reduces the cluster iteration to fall into the local extremum the possibility. Regarding color image's division, discussed the color space selection which the cluster uses, determined that the YUV color space takes the EFCM algorithm division color image the color space, as well as in view of YUV color space based on brightness and chromatic aberration Sobel marginal check operator.