Robust Tucker Tensor Decomposition for Effective Image Representation

Many tensor based algorithms have been proposed for the study of high dimensional data in a large variety of computer vision and machine learning applications. However, most of the existing tensor analysis approaches are based on Frobenius norm, which makes them sensitive to outliers, because they minimize the sum of squared errors and enlarge the influence of both outliers and large feature noises. In this paper, we propose a robust Tucker tensor decomposition model (RTD) to suppress the influence of outliers, which uses L1-norm loss function. Yet, the optimization on L1-norm based tensor analysis is much harder than standard tensor decomposition. In this paper, we propose a simple and efficient algorithm to solve our RTD model. Moreover, tensor factorization-based image storage needs much less space than PCA based methods. We carry out extensive experiments to evaluate the proposed algorithm, and verify the robustness against image occlusions. Both numerical and visual results show that our RTD model is consistently better against the existence of outliers than previous tensor and PCA methods.

[1]  Yi Ma,et al.  Robust principal component analysis? , 2009, JACM.

[2]  Allen Y. Yang,et al.  A Review of Fast L(1)-Minimization Algorithms for Robust Face Recognition , 2010 .

[3]  Michael J. Black,et al.  A Framework for Robust Subspace Learning , 2003, International Journal of Computer Vision.

[4]  P. N. Bellhumer Eigenfaces vs. fisherfaces : Recognition using class specific linear projection , 1997 .

[5]  Junbin Gao,et al.  Robust L1 Principal Component Analysis and Its Bayesian Variational Inference , 2008, Neural Computation.

[6]  Kohei Inoue,et al.  Equivalence of Non-Iterative Algorithms for Simultaneous Low Rank Approximations of Matrices , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[7]  Amnon Shashua,et al.  Linear image coding for regression and classification using the tensor-rank principle , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[8]  Demetri Terzopoulos,et al.  Multilinear Analysis of Image Ensembles: TensorFaces , 2002, ECCV.

[9]  Ying Wu,et al.  Locality Versus Globality: Query-Driven Localized Linear Models for Facial Image Computing , 2008, IEEE Transactions on Circuits and Systems for Video Technology.

[10]  L. Tucker,et al.  Some mathematical notes on three-mode factor analysis , 1966, Psychometrika.

[11]  Chris H. Q. Ding,et al.  Are Tensor Decomposition Solutions Unique? On the Global Convergence HOSVD and ParaFac Algorithms , 2009, PAKDD.

[12]  M. Turk,et al.  Eigenfaces for Recognition , 1991, Journal of Cognitive Neuroscience.

[13]  R. Tibshirani,et al.  Sparse Principal Component Analysis , 2006 .

[14]  D. Hawkins,et al.  Methods of L1 estimation of a covariance matrix , 1987 .

[15]  Chris H. Q. Ding,et al.  2-Dimensional Singular Value Decomposition for 2D Maps and Images , 2005, SDM.

[16]  Dimitri P. Bertsekas,et al.  Nonlinear Programming , 1995 .

[17]  W. Krzanowski,et al.  A Characterization of Principal Components for Projection Pursuit , 1999 .

[18]  Allen Y. Yang,et al.  A Review of Fast l1-Minimization Algorithms for Robust Face Recognition , 2010, ArXiv.

[19]  Jieping Ye,et al.  Generalized Low Rank Approximations of Matrices , 2004, Machine Learning.

[20]  Yuxiao Hu,et al.  Learning a locality preserving subspace for visual recognition , 2003, Proceedings Ninth IEEE International Conference on Computer Vision.

[21]  Alejandro F. Frangi,et al.  Two-dimensional PCA: a new approach to appearance-based face representation and recognition , 2004 .

[22]  Joos Vandewalle,et al.  A Multilinear Singular Value Decomposition , 2000, SIAM J. Matrix Anal. Appl..

[23]  Philippe C. Besse,et al.  A L 1-norm PCA and a Heuristic Approach , 1996 .

[24]  Takeo Kanade,et al.  Robust L/sub 1/ norm factorization in the presence of outliers and missing data by alternative convex programming , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[25]  Nojun Kwak,et al.  Principal Component Analysis Based on L1-Norm Maximization , 2008, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[26]  David J. Kriegman,et al.  Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection , 1996, ECCV.

[27]  Dimitri P. Bertsekas,et al.  Nonlinear Programming 2 , 2005 .