This paper proposes a new algorithm, multi-task robust principal component analysis (MTRPCA),to collaboratively integrate multiple visual features for motion segmentation. Given the video data described by multiple features, the motion parts are obtained by jointly decomposing multiple feature matrices into pairs of low-rank and sparse matrices. The inference process is formulated as a convex optimization problem that minimizes a constrained combination of nuclear norm and `2,1-norm. The convex optimization problem can be solved efficiently with an augmented Lagrange multiplier (ALM) method. Compared with previous methods based on individual features, the proposed method seamlessly integrates multiple features within a single inference
step, and thus produces more accurate and reliable results. Experiments on human motion data sets, Human Eva and change detection, show that the proposed MTRPCA is effective and promising.
WANG Xiang-yang1,2, WAN Wang-gen1,2
. Motion Segmentation via Multi-task Robust Principal Component Analysis[J]. Journal of Applied Sciences, 2014
, 32(5)
: 473
-480
.
DOI: 10.3969/j.issn.0255-8297.2014.05.007
[1] SIGAL Leonid, BALAN Alexandru O, BLACK Michael J. HumanEva: synchronized video and motion capture dataset and baseline algorithm for evaluation of articulated human motion [J]. International Journal of Computer Vision, 2010, 87: 4-27, 2010.
[2] BOUWMANS T, BAF F E, VACHON B. Statistical background modeling for foreground detection: a survey [J]. Handbook of Pattern Recognition and Computer Vision, World Scientific Publishing, 2010, 4(2): 181-189.
[3] BOUWMANS T. Subspace learning for background modeling: a survey. Recent Patents on Computer Science, 2009, 2(3): 223-234.
[4] TORRE F D L, BLACK M. A framework for robust subspace learning [J]. International Journal on Computer Vision, 2003: 117-142.
[5] WRIGHT John, PENG Yigang, MA Yi, GANESH Arvind, RAO Shankar. Robust principal component analysis: exact recovery of corrupted low-rank matrices by convex optimization [C]//Proceedings of Advances in Neural Information Processing Systems, December 2009 (NIPS 2009).
[6] CANDÈS Emmanuel J, LI X, MA Y, WRIGHT John. Robust principal component analysis [J]. Journal of the ACM, 2011, 58 (3): 1-37.
[7] LIN Zhouchen, CHEN Minming, WU Leqin, MA Yi. The augmented Lagrange multiplier method for exact recovery of corrupted low-rank matrices [R]. UIUC Technical Report UILU-ENG-09-2215, 2009.
[8] BECKER S, CANDES E, GRANT M. Tfocs. Flexible first-order methods for rank minimization. Low-rank Matrix Optimization Symposium, SIAM Conference on Optimization, 2011.
[9] DING Xinghao, HE Lihan, CARIN Lawrence. Bayesian robust principal component analysis[J]. IEEE Transaction on Image Processing, 2011, 20(12): 3419- 3430.
[10] LIU Guangcan, LIN Zhouchen, YU Yong. Robust subspace segmentation by low-rank representation [C]//The 27th International Conference on Machine Learning (ICML2010), June 21-24, 2010, Haifa, Israel.
[11] LIU Guangcan, LIN Zhouchen, YAN Shuicheng, SUN Ju, YU Yong, MA Yi. Robust recovery of subspace structures by low-rank representation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence(TPAMI), 2013, 35(1): 171-184.
[12] CHENG Bin, LIU Guangcan, WANG Jingdong, HUANG Zhongyang, YAN Shuicheng. Multi-task low-rank affinities pursuit for image segmentation [C]//13th International Conference on Computer Vision (ICCV2011), November 6-13, 2011, Barcelona, Spain.
[13] LANG Congyan, LIU Guangcan, YU Jian, YAN Shuicheng. Saliency detection by multitask sparsity pursuit [J]. IEEE Transactions on Image Processing, 2012, 21(3): 1327-1338.
[14] CAI Jian Feng, CANDES Emmanuel J, SHEN Zuo Wei. A singular value thresholding algorithm for matrix completion [J]. SIAM Journal on Optimization, 2010, 20(4): 1956-1982.
[15] GOYETTE N, JODOIN P M, PORIKLI F, KONRAD J, ISHWAR P. Changedetection.net: a new change detection benchmark dataset [C]//Proc. IEEE Workshop on Change Detection (CDW-12) at CVPR 2012, Providence, RI, 16-21 June, 2012
[16] KAMARAINEN Joni-Kristian, KYRKI Ville, KALVIAINEN Heikki. Invariance properties of Gabor filter-based features-overview and applications [J]. IEEE Transactions on Image Processing, 2006, 15(5): 1088-1099.
[17] MADDALENA L, PETROSINO A. A fuzzy spatial coherence-based approach to background foreground separation for moving object detection [J]. Neural Computing and Applications, 2010: 1-8.
[18] GUYON C, BOUWMANS T, ZAHZAH E. Robust principal component analysis for background subtraction: systematic evaluation and comparative analysis [M]. Principal Component Analysis, Book 1, Chapter 12, 2012: 223-238.
