Tensor Completion Algorithm and Its Applications in Face Recognition
SHI Jia-Rong, JIAO Li-Cheng, SHANG Fan-Hua
Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education of China, Institute of Intelligent Information Processing, Xidian University, Xian 710071
Abstract Missing data problems are commonly attributed to the matrix completion problem, and matrix completion is an important method of signal acquisitions following compressing sensing. The data examples have the property of multi-linearity in applications, that is, the data set can be represented by higher order tensors. The tensor completion problem and its applications in face recognition are studied. Based on lower-dimensional Tucker decomposition of tensors, an iterative algorithm is proposed to complete tensors. And the distance between the estimating tensor and its Tucker approximation tensor monotonically decreases during the iterative procedure. Experimental results demonstrate the effectiveness and feasibility of the proposed method in completing tensor and face recognition.
[1] Okatani T, Deguchi K. On the Wiberg Algorithm for Matrix Factorization in the Presence of Missing Components. International Journal of Computer Vision, 2007, 72(3): 329-337 [2] Chen Pei. Optimization Algorithms on Subspaces: Revisiting Missing Data Problem in Low-Rank Matrix. International Journal of Computer Vision, 2008, 80(1): 125-142 [3] Korah T, Rasmussen C. Spatiotemporal Inpainting for Recovering Texture Maps of Occluded Building Facades. IEEE Trans on Image Processing, 2007, 16(9): 2262-2271 [4] Garcia-Laencina P J, Sancho-Gomez J L, Figueiras-Vidal A R, et al. K Nearest Neighbours with Mutual Information for Simultaneous Classification and Missing Data Imputation. Neurocomputing, 2009, 72(7/8/9): 1483-1493 [5] Farhangfara A, Kurganb L, Dy J. Impact of Imputation of Missing Values on Classification Error for Discrete Data. Pattern Recognition, 2008, 41(12): 3692-3705 [6] Vidal R, Tron R , Hartley R. Multiframe Motion Segmentation with Missing Data Using Power Factorization and GPCA. International Journal of Computer Vision, 2008, 79(1): 85-105 [7] Candès E J, Recht B. Exact Matrix Completion via Convex Optimization. Foundations of Computational Mathematics, 2009, 9(6): 717-772 [8] Candès E J, Terence T. The Power of Convex Relaxation: Near-Optimal Matrix Completion. IEEE Trans on Information Theory, 2010, 56(5): 2053-2080 [9] Cai Jianfeng, Candès E J, Shen Zouwei. A Singular Value Thresholding Algorithm for Matrix Completion. SIAM Journal on Optimization, 2010, 20(4): 1956-1982 [10] Candès E J, Plan Y. Matrix Completion with Noise. Proc of the IEEE, 2010, 98(6): 925-936 [11] Yan Hui, Jin Zhong, Yang Jingyu. Non-Negative Two-Dimensional Principal Component Analysis and Its Application to Face Recognition. Pattern Recognition and Artificial Intelligence, 2009, 22(6): 809-814 (in Chinese) (严 慧,金 忠,杨静宇.非负二维主成分分析及在人脸识别中的应用.模式识别与人工智能, 2009, 22(6): 809-814) [12] Wang Hongcheng, Ahuja N. A Tensor Approximation Approach to Dimensionality Reduction. International Journal of Computer Vision, 2008, 76(3): 217-229 [13] Yan Shuicheng, Xu Dong, Yang Qiang, et al. Multilinear Discriminant Analysis for Face Recognition. IEEE Trans on Image Processing, 2007, 16(1): 212-220 [14] Kolda T G, Bader B W. Tensor Decompositions and Applications. SIAM Review, 2009, 51(3): 455-500 [15]Lathauwer L D, Moor B D, Vandewalle J. A Multilinear Singular Value Decomposition. SIAM Journal on Matrix Analysis and Applications, 2000, 21(4): 1253-1278 [16] Lathauwer L D, Moor B D, Vandewalle J. On the Best Rank-1 and Rank-(R1,R2,…,RN) Approximation of Higher-Order Tensors. SIAM Journal on Matrix Analysis and Applications, 2000, 21(4): 1324-1342 [17] Tomasi G, Bro R. PARAFAC and Missing Values. Chemometrics and Intelligent Laboratory Systems, 2005, 75(2): 163-180 [18] Liu Ji, Musialski P, Wonka1 P, et al. Tensor Completion for Estimating Missing Values in Visual Data // Proc of the 20th IEEE International Conference on Computer Vision. Kyoto, Japan, 2009: 2114-2121
2011, Vol. 24 
