Automatic Spectral Clustering and its Application

An new algorithm called automatic spectral clustering (ASC) is proposed based on eigengap and orthogonal eigenvector in this paper. It mainly focuses on how to automatically determine the suitable class number in clustering and explores some intrinsic characteristics of the spectral clustering method. The proposed method firstly constructs the affinity matrix of data and carries on eigen-decomposition, then determine the class number according to the eigengap. Finally, the data are classified by employing the angle between two eigenvectors. The experiments on the real–world data sets from UCI and applications in face location show the correctness and efficiency of the proposed method.

[1]  Ding Lin,et al.  Image segmentation using spectral clustering , 2005, 17th IEEE International Conference on Tools with Artificial Intelligence (ICTAI'05).

[2]  Ulrike von Luxburg,et al.  A tutorial on spectral clustering , 2007, Stat. Comput..

[3]  Kun Huang,et al.  A unifying theorem for spectral embedding and clustering , 2003, AISTATS.

[4]  Martine D. F. Schlag,et al.  Spectral K-way ratio-cut partitioning and clustering , 1994, IEEE Trans. Comput. Aided Des. Integr. Circuits Syst..

[5]  James C. Bezdek,et al.  Pattern Recognition with Fuzzy Objective Function Algorithms , 1981, Advanced Applications in Pattern Recognition.

[6]  Mikhail Belkin,et al.  Consistency of spectral clustering , 2008, 0804.0678.

[7]  Li Xiao.  Spectral clustering based on matrix perturbation theory , 2007 .

[8]  Shanan Zhu,et al.  Multi-face detection based on downsampling and modified subtractive clustering for color images , 2007 .

[9]  Michael I. Jordan,et al.  On Spectral Clustering: Analysis and an algorithm , 2001, NIPS.

[10]  Tian Zheng,et al.  Spectral clustering based on matrix perturbation theory , 2007 .