Article Contents
Article Contents

# Multi-view foreground segmentation via fourth order tensor learning

• In this paper, we present a novel fuse-before-detect algorithm for multi-view foreground segmentation via fourth order tensor learning. By using several camera views, most of the existing algorithms first detect the various object features for each view and then fuse the data together for foreground segmentation or tracking. However, this kind of single view foreground segmentation algorithm always suffers from various environmental problems, such as reflection and shadow induced by shiny objects, especially floor and wall. These segmentation errors reduce the accuracy of the multi-view tracking algorithms. In the proposed algorithm, we first fuse multi-view camera data to a fourth-order tensor through multiple parallelized planes projections. An incremental fourth-order tensor learning algorithm is then employed to perform foreground segmentation in the fused tensor data. By collecting all the information from different views, this approach could restrain the specific environmental effects in each view and give better segmentation results. Experimental results are reported to show the performance of the proposed method is better than the state-of-the-art methods in challenged environments.
Mathematics Subject Classification: Primary: 15A18, 15A69, 65F25, 65F30.

 Citation:

•  [1] C. Stauffer and W. E. L. Grimson, Adaptive background mixture models for real-time tracking, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2 (1999), 246-252.doi: 10.1109/CVPR.1999.784637. [2] A. Elgammal, D. Harwood and L. Davis, Non-parametric model for background subtraction, Proceedings of the European Conference on Computer Vision, 1843 (2000), 751-767. [3] W. Hu, X. Li, X. Zhang, X. Shi, S. Maybank and Z. Zhang, Incremental tensor subspace learning and its applications to foreground segmentation and tracking, International Journal of Computer vision, 91 (2011), 303-327. [4] M. Taj and A. Cavallaro, Multi-view multi-object detection and tracking, Computer Vision, 285 (2010), 263-280.doi: 10.1007/978-3-642-12848-6_10. [5] S. M. Khan and M. Shah, Tracking multiple occluding people by localizing on multiple scene planes, IEEE Transaction on Pattren Analysis and Machine Intelligence, 31 (2009), 505-519.doi: 10.1109/TPAMI.2008.102. [6] A. Criminisi, I. Reid and A. Zisserman, Single view metrology, International Journal of Computer Vision, 40 (2000), 123-148. [7] L. De Lathauwer, B. De Moor and J. Vandewalle, On the best rank-1, and rank-($R_1$, $R_2$, ..., $R_n$) approximation of higher-order tensors, SIAM Journal of Matrix Analysis and Applications, 21 (2000), 1324-1342.doi: 10.1137/S0895479898346995. [8] R. Hartley and A. Zisserman, "Multiple View Geometry in Computer Vision," Second edition, Cambridge Univ. Press, Cambridge, 2003. [9] D. A. Ross, J. Lim, R.-S. Lin and M.-H. Yang, Incremental learning for robust visual tracking, International Journal of Computer Vision, 77 (2008), 125-141.doi: 10.1007/s11263-007-0075-7. [10] B. Li, K. Peng, X. Ying and H. Zha, Simultaneous vanishing point detection and camera calibration from single images, in "Advances in Visual Computing," Lecture Notes in Computer Science, 6454, Springer, (2010), 151-160.doi: 10.1007/978-3-642-17274-8_15.