Abstract: Aiming at the problems of the sensitivity to noise,the low computational efficiency and the segmentation result depended on heavily the initial value,we put forward an image segmentation method based on KFCM and improved CV model.Firstly,the average nuclear fuzzy clustering method was used to determine the interest area as the initial value in the algorithm.Then we improved the efficiency of iterative calculation of the CV model in calculation and time by the Split Bregman method.The experimental results show that this algorithm not only keeps the advantage of image segmentation algorithm,but also has obvious effect on the noise performance and segmentation efficiency．

Key words: KFCM agorithm,CV model,Image segmentation,Split Bregman method