Visual tracking based on the sparse representation of the PCA subspace

Abstract

We construct a collaborative model of the sparse representation and the subspace representation. First, we represent the tracking target in the principle component analysis (PCA) subspace, and then we employ an L₁ regularization to restrict the sparsity of the residual term, an L₂ regularization term to restrict the sparsity of the representation coefficients, and an L₂ norm to restrict the distance between the reconstruction and the target. Then we implement the algorithm in the particle filter framework. Furthermore, an iterative method is presented to get the global minimum of the residual and the coefficients. Finally, an alternative template update scheme is adopted to avoid the tracking drift which is caused by the inaccurate update. In the experiment, we test the algorithm on 9 sequences, and compare the results with 5 state-of-art methods. According to the results, we can conclude that our algorithm is more robust than the other methods.

References

[1] WangD, YangJ, YangL, GeH. Journal of Optoelectronics ·Laser, 2016, 27: 1066

[2] LiC, ShiJ, XuL, WeiL. Journal of Optoelectronics ·Laser, 2015, 26: 1381

[3] YangH, WuX, HeB, ZhuM. Journal of Optoelectronics ·Laser, 2015, 26: 170

[4] BaiT, LiY. IEEE Trans. on Industrial Electronics, 2014, 10: 538

[5] RossD A, LimJ, LinR S, YangM H. International Journal of Computer Vision, 2008, 77: 125

[6] TurkM A, PentlandA P. J. Cogn. Neurosci., 2001, 3: 71

[7] ShiraziS, HarandiM, LovellBC, SandersonC. Object Tracking via Non-euclidean Geometry: A Grassmann Approach, Proc. IEEE Winter Conf. on Applications of Comput. Vision (WACV), 2014, 901

[8] KumarBS, SwamyMNS, AhmadMO. Weighted Residual Minimization in PCA Subspace for Visual Tracking, IEEE International Symposium on Circuits and Systems, 2016, 986

[9] WrightJ, YangA, GaneshA, SastryS, MaY. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2008, 31: 210

[10] YangJ, WrightJ, HuangT S, MaY. IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society, 2010, 19: 2861

[11] CevherV, SankaranarayananA, DuarteM F, ReddyD, BaraniukR G, ChellappaR. Compressive Sensing for Background Subtraction, European Conference on Computer Vision, 2008, 155

[12] MeiX, LingH B, WuY, BlaschE, BaiLMinimum Error Bounded Efficient L1 Tracker with Occlusion DetectionThe 24th IEEE Conference on Computer Vision and Pattern Recognition, 2011, 1257

[13] MeiX, LingH BRobust Visual Tracking using l1 MinimizationIEEE International Conference on Computer Vision, 2009, 1436

[14] BaoC L, WuY, LingH B, JiHReal Time Robust L1 Tracker using Accelerated Proximal Gradient Approach, IEEE Conference on Computer Vision and Pattern RecognitionIEEE Computer Society, 2012, 1830

[15] ZhuangBH, LuHC, XiaoZY, WangD. IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society, 2014, 23: 1872

[16] WangD, LuHC, YangMHLeast Soft-threshold Squares TrackingIEEE Computer Vision and Pattern Recognition, 2013, 2371

[17] XiaoZY, LuHC, WangD. IEEE Transactions on Circuits & Systems for Video Technology, 2014, 24: 1301

[18] WuY, LimJ, YangMH. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 37: 1834

[19] ZhongW, LuHC, YangMH. Robust Object Tracking via Sparsity-based Collaborative Model, IEEE Computer Vision and Pattern Recognition, 2012, 1838

[20] KalalZ, MikolajczykK, MatasJR. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2012, 34: 1409

[21] ZhangT, GhanemB, LiuS. International Journal of Computer Vision, 2013, 101: 367

[22] EveringhamM, Van GoolL, WilliamsCK, WinnJ, ZissermanA. International Journal of Computer Vision, 2010, 88: 303