Abstract: There could be some information loss when the traditional median filter alone is applied to remove the impulse noise in images. To overcome this drawback,a novel iterative median filtering algorithm based on the improved spring-mass model is proposed. Firstly,the detected point is considered as a center such that the pulling forces from the eight neighboring pixels form a spring-mass model in plane. The stationary condition of the spring-mass model is used to determine whether the detected point is a noisy point. Secondly,by using iterative method,we use only signal points to modify the pixel values of the noisy points. The experimental results show that the proposed algorithm is effective in noise suppression and detail preservation compared with conventional methods.