摘要
传统基于模糊C均值聚类图像分割算法易受复杂纹理和噪声干扰,无法准确分割图像。针对这一现象,提出一种基于权重系数模糊C均值聚类算法,并将其应用于图像分割中。算法定义权重系数矩阵,将每个像点的邻域信息引入到像点间相似性度量中,计算每个像点与聚类中心点的邻域相似程度,根据权重系数矩阵确定邻域中每个像点在邻域特征计算中所占权重,增强了算法对噪点和杂波的鲁棒性。实验结果表明,与传统模糊C均值聚类算法相比,该文算法获得更加精确的图像分割结果。
A weighting coefficient matrix based fuzzy C--means clustering algorithm for image segmentation was proposed to solve the problems that the segmentation results of the traditional FCM based image segmentation algorithms were easily disturbed by complex texture and noise. In this algorithm,weighting coefficient matrix was defined to calculate neighborhood feature for every pixel in the image,and neighborhood information for every pixel in the image was introduced into similarity measure calculation between pixels and cluster centers,that can improve the robustness of the improved algorithm to noise and clutter. The experimental results demonstrated that the propnsed algorithm achieves more accurate image segmentation compared with traditional FCM algorithms.
出处
《科技资讯》
2015年第11期245-247,共3页
Science & Technology Information
