摘要
提出一种基于二维直方图加权的模糊c均值图像快速分割算法.通过将原图像和它的平滑图像相结合,构造一个二元组的“广义图像”,广义图像的直方图就是原图像的二维直方图.然后对此二维直方图进行塔形分解得到金字塔的上一层——顶层,相应地称原二维直方图为底层.最后,利用加权模糊c均值聚类算法分别对顶层和底层进行模糊聚类,从而实现对原图像的分割.实验结果与性能分析表明,该算法具有较高的分割速度和良好的抑制噪声的能力.
This paper studies the application of fuzzy c-means （FCM） clustering algorithm in the image segmentation, and a fast image segmentation method is presented based on a 2D histogram weighting FCM algorithm. Firstly, by combining original image and its smooth image, a generalized image is constructed. Secondly, the 2D histogram is decomposed into a pyramid structure with 2 layers（one upper and one lower）. Finally, the weighting FCM algorithm is performed on the two layers respectively to implement the original image segmentation. The experimental results illustrate that the pro- posed algorithm can implement image segmentation quickly and suppress the noise effectively.
出处
《西安建筑科技大学学报（自然科学版）》
CSCD
北大核心
2007年第2期280-285,共6页
Journal of Xi'an University of Architecture & Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(60202004)
教育部重点项目(104173)
关键词
图像分割
加权模糊c均值聚类算法
塔形结构
二维直方图
image segmentation
weighting fuzzy c-means clustering algorithm
pyramid structure
2D histogram
