关键词: 抑制噪音, 模糊C均值聚类算法, 塔形结构, 图像分割

Abstract: An effective noise reduction fuzzy C-means clustering algorithm is proposed.It reduces the impact of noise on cluster centers by constructing gray-median-based spatial information and pyramid structure.The pyramid structure is introduced to speed up the approach and the adaptive threshold selection of membership grade avoids the inflexibility of human threshold.Median images instead of original images are used to eliminate noise.Simulation results show that this method has a strong robustness for noise image and the segmentation results are more accurate.

Key words: noise reduction, fuzzy C-means clustering algorithm, pyramid structure, image segmentation