Constrained Nonnegative Matrix Factorization for Image Representation

被引:426
作者
Liu, Haifeng [1 ]
Wu, Zhaohui [1 ]
Li, Xuelong [2 ]
Cai, Deng [3 ]
Huang, Thomas S. [4 ]
机构
[1] Zhejiang Univ, Coll Comp Sci, Hangzhou 310027, Zhejiang, Peoples R China
[2] Chinese Acad Sci, Ctr Opt IMagery Anal & Learning OPTIMAL, State Key Lab Transient Opt & Photon, Xian Inst Opt & Precis Mech, Xian 710119, Shaanxi, Peoples R China
[3] Zhejiang Univ, State Key Lab CAD&CG, Coll Comp Sci, Hangzhou 310058, Zhejiang, Peoples R China
[4] Univ Illinois, Dept Elect & Comp Engn, Beckman Inst Adv Sci & Technol, Urbana, IL 61801 USA
基金
中国国家自然科学基金;
关键词
Nonnegative matrix factorization; semi-supervised learning; dimension reduction; clustering; GRADIENT; PARTS;
D O I
10.1109/TPAMI.2011.217
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Nonnegative matrix factorization (NMF) is a popular technique for finding parts-based, linear representations of nonnegative data. It has been successfully applied in a wide range of applications such as pattern recognition, information retrieval, and computer vision. However, NMF is essentially an unsupervised method and cannot make use of label information. In this paper, we propose a novel semi-supervised matrix decomposition method, called Constrained Nonnegative Matrix Factorization (CNMF), which incorporates the label information as additional constraints. Specifically, we show how explicitly combining label information improves the discriminating power of the resulting matrix decomposition. We explore the proposed CNMF method with two cost function formulations and provide the corresponding update solutions for the optimization problems. Empirical experiments demonstrate the effectiveness of our novel algorithm in comparison to the state-of-the-art approaches through a set of evaluations based on real-world applications.
引用
收藏
页码:1299 / 1311
页数:13
相关论文
共 38 条
[1]  
[Anonymous], 2003, INT C MACH LEARN
[2]  
[Anonymous], 2001, P ADV NEUR INF PROC
[3]  
[Anonymous], P IEEE INT C COMP VI
[4]  
[Anonymous], 1973, Pattern Classification and Scene Analysis
[5]  
[Anonymous], 2003, ADV NEURAL INFORM PR
[6]  
Belkin M, 2006, J MACH LEARN RES, V7, P2399
[7]  
BUNTINE W, 2002, P 13 EUR C MACH LEAR
[8]   Non-negative Matrix Factorization on Manifold [J].
Cai, Deng ;
He, Xiaofei ;
Wu, Xiaoyun ;
Han, Jiawei .
ICDM 2008: EIGHTH IEEE INTERNATIONAL CONFERENCE ON DATA MINING, PROCEEDINGS, 2008, :63-+
[9]   MAXIMUM LIKELIHOOD FROM INCOMPLETE DATA VIA EM ALGORITHM [J].
DEMPSTER, AP ;
LAIRD, NM ;
RUBIN, DB .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-METHODOLOGICAL, 1977, 39 (01) :1-38
[10]  
Ding C., 2006, P AAAI C ART INT