关键词: 模糊C均值聚类, 相关矩阵, 脉冲噪声, 中值滤波, 图像降噪, 聚类分离

Abstract: In order to improve the filtering effect of noise images, a new filtering algorithm is proposed. The obvious noise of the noise image is distinguished in the preliminary filtering. The correlation matrix of the noise image is calculated according to the similarities of the local pixels and the isolations of the noise points, and then noise points and normal pixel are  separated by using the fuzzy c-means algorithm to cluster the correlation matrix iteratively. The noise points are filtered with median filter algorithm. Simulation results indicate that this algorithm can not only accurately filter more noise points and protect more image details, but also get better filtering effect than traditional algorithms mentioned in this paper.

Key words: fuzzy c-means algorithm clustering, related matrix, impulsive noise, median filtering, image noise reduction, clustering separation