基于粗糙集理论和FCM的图像聚类方法

Image Clustering Based on Rough Set Theory and FCM

提出一种将粗糙集方法与模糊C均值聚类(FCM)算法结合的图像聚类方法。借助于粗糙集理论在处理大数据量、消除冗余信息等方面的优点,减少模糊C均值聚类的训练数据量,克服其因为数据量大而处理速度慢等缺点,同时利用模糊C均值聚类好的聚类性能,对经过约简的最小属性子集进行聚类分析,实现图像聚类的快速、准确、鲁棒等优点。在人脸图像上的聚类实验取得了很好的效果。

An image clustering method which combines rough set theory and fuzzy C-mean clustering is proposed. In virtue of the advantage of rough set theory, such as large training data processing and redundant information elimination, the method can reduce the operation amount of the fuzzy C-mean clustering and overcome the defect of slow training speed when the training data are enor- mous. Clustering analysis is carried on rapidly, exactly, robustly in the reduced core of property set by fuzzy C-mean clustering. The experiment in the face image demonstrates the validity.