Computer Engineering and Applications ›› 2010, Vol. 46 ›› Issue (5): 184-185.DOI: 10.3778/j.issn.1002-8331.2010.05.056

• 图形、图像、模式识别 • Previous Articles     Next Articles

Weighted semi-fuzzy C-means based on histograms

YANG Dao-pu,MA Qiu-he,SHI Lei,CHEN Ke   

  1. Institute of Surveying and Mapping,Information Engineering University,Zhengzhou 450052,China
  • Received:2008-09-22 Revised:2008-12-10 Online:2010-02-11 Published:2010-02-11
  • Contact: YANG Dao-pu


杨道普,马秋禾,石 磊,陈 科   

  1. 解放军信息工程大学 测绘学院,郑州 450052
  • 通讯作者: 杨道普

Abstract: To consider the feature of image data,this paper proposes a weighted semi-fuzzy C-means algorithm.The weight is derived from histograms of image.The method also takes in HCM’s advantage.Considering the local and the entire information,it acquires a better effect.Compared with traditional methods,the test experiment with Lena and brain image demonstrates effectiveness of the algorithm,and the experiment of processing a larger image is given to show the method is faster.

Key words: histograms, Fuzzy C-Means clustering(FCM), Hard C-Means clustering(HCM), semi-WFCM, Otsu, image segment

摘要: 通过分析影像数据的特点,利用直方图的统计特性,结合HCM收敛速度快的优点,提出了一种基于直方图加权的半模糊化的聚类算法,此方法结合了全局与局部信息,提高了聚类的速度,改善了聚类的效果;采用Lena和脑影像实验与传统算法作比较证明了该算法的效果更好,并对一副97 658k的影像进行处理,证明了该算法效率高。

关键词: 直方图, 模糊C-均值聚类算法(FCM), 硬C-均值聚类算法(HCM), 半模糊加权聚类, Otsu, 影像分割

CLC Number: