Exploiting a depth contextmodel in visual tracking with correlation filter

Abstract

Recently correlation filter based trackers have attracted considerable attention for their high computational efficiency. However, they cannot handle occlusion and scale variation well enough. This paper aims at preventing the tracker from failure in these two situations by integrating the depth information into a correlation filter based tracker. By using RGB-D data, we construct a depth context model to reveal the spatial correlation between the target and its surrounding regions. Furthermore, we adopt a region growing method to make our tracker robust to occlusion and scale variation. Additional optimizations such as a model updating scheme are applied to improve the performance for longer video sequences. Both qualitative and quantitative evaluations on challenging benchmark image sequences demonstrate that the proposed tracker performs favourably against state-of-the-art algorithms.

Keywords

Visual tracking / Depth context model / Correlation filter / Region growing

References

[1] Adam,A., Rivlin, E., Shimshoni,I. , 2006. Robust fragmentsbased tracking using the integral histogram. IEEE Computer Society Conf. on Computer Vision and Pattern Recognition, p.798–805.

[2] Adams,R., Bischof, L., 1994. Seeded region growing. IEEE Trans. Patt. Anal. Mach. Intell., 16(6):641–647.

[3] Bolme,D.S., Beveridge, J.R., Draper,B.A. , , 2010. Visual object tracking using adaptive correlation filters. IEEE Conf. on Computer Vision and Pattern Recognition, p.2544–2550.

[4] Cehovin,L., Kristan, M., Leonardis,A. , 2011. An adaptivecoupled-layer visual model for robust visual tracking. IEEE Int. Conf. on Computer Vision, p.1363–1370.

[5] Chen,K., Lai,Y., Wu,Y., , 2014. Automatic semantic modeling of indoor scenes from low-quality RGB-D datausing contextual information. ACM Trans. Graph., 33(6):208–219.

[6] Choi,C., Christensen, H.I., 2013. RGB-D object tracking: aparticle filter approach on GPU. IEEE/RSJ Int. Conf. on Intelligent Robots and Systems, p.1084–1091.

[7] Danelljan,M., Häger, G., Khan,F.S. , , 2014a. Accuratescale estimation for robust visual tracking. British Machine Vision Conf., p.1–11.

[8] Danelljan,M., Khan,F.S., Felsberg,M. , , 2014b. Adaptivecolor attributes for real-time visual tracking. IEEE Conf. on Computer Vision and Pattern Recognition, p.1090–1097.

[9] Dinh,T.B., Vo,N., Medioni,G.G. , 2011. Context tracker: exploring supporters and distracters in unconstrained environments. IEEE Conf. on Computer Vision and Pattern Recognition, p.1177–1184.

[10] Everingham,M., Gool,L.V., Williams,C.K. , , 2010. The Pascal Visual Object Classes (VOC) challenge. Int. J. Comput. Vis., 88(2):303–338.

[11] Grabner,H., Matas, J., Gool,L.V. , , 2010. Tracking theinvisible: learning where the object might be. IEEE Conf. on Computer Vision and Pattern Recognition, p.1285–1292.

[12] Gupta,S., Girshick, R.B., Arbelaez,P. , , 2014. Learning rich features from RGB-D images for object detection and segmentation. ECCV, p.345–360.

[13] Hare,S., Saffari, A., Torr,P. , , 2011. Struck: structured output tracking with kernels. IEEE Trans. Patt. Anal. Mach. Intell., 38(10):263–270.

[14] Henriques,J.F., Caseiro, R., Martins,P. , , 2012. Exploitingthe circulant structure of tracking-by-detectionwith kernels. ECCV, p.702–715.

[15] Henriques,J.F., Caseiro, R., Martins,P. , , 2015. Highspeed tracking with kernelized correlation filters. IEEE Trans. Patt. Anal. Mach. Intell., 37(3):583–596.

[16] Hickson,S., Birchfield, S., Essa,I.A. , , 2014. Efficient hierarchical graph-based segmentation of RGBD videos. IEEE Conf. on Computer Vision and Pattern Recognition, p.344–351.

[17] Izadinia,H., Saleemi, I., Li,W. , , 2012. (MP) 2T:multiple people multiple parts tracker. IEEE Conf. on Computer Vision and Pattern Recognition, p.100–114.

[18] Kalal,Z., Mikolajczyk, K., Matas,J. , 2012. Trackinglearning-detection. IEEE Trans. Patt. Anal. Mach. Intell., 34(7):1409–1422.

[19] Kristan,M., Pflugfelder, R., Leonardis,A. , , 2015. Thevisual object tracking VOT2014 challenge results.IEEE Conf. on Computer Vision and Pattern Recognition, p.191–217.

[20] Kumar,B.V., Mahalanobis, A., Juday,R.D. , 2010. Correlation Pattern Recognition. Cambridge University Press, Cambridge.

[21] Lee,D., Sim,J., Kim,C., 2014. Visual tracking using pertinent patch selection and masking. IEEE Conf. on Computer Vision and Pattern Recognition, p.3486–3493.

[22] Li,X., Hu,W., Shen,C., , 2013. A survey of appearance models in visual object tracking. ACM Intell. Syst. Technol., 4(4):58.

[23] Li,Y., Zhu,J., 2014. A scale adaptive kernel correlation filter tracker with feature integration. ECCV, p.254–265.

[24] Li,Y., Zhu,J., Hoi,S., , 2015. Reliable patch trackers: robust visual tracking by exploiting reliable patches. IEEE Conf. on Computer Vision and Pattern Recognition, p.353–361.

[25] Liu,T., Wang,G., Yang,Q., 2015. Real-time part-basedvisual tracking via adaptive correlation filters. IEEE Conf. on Computer Vision and Pattern Recognition, p.4902–4912.

[26] Luber,M., Spinello, L., Arras,K.O. , 2011. People trackingin RGB-D data with on-line boosted target models. IEEE/RSJ Int. Conf. on Intelligent Robots and Systems, p.3844–3849.

[27] Ma,C., Yang,X., Zhang,C., , 2015. Long-termcorrelation tracking. IEEE Conf. on Computer Visionand Pattern Recognition, p.5388–5396.

[28] Park,Y., Lepetit, V., Woo,W. , 2011. Texture-less object tracking with online training using an RGB-D camera.10th IEEE Int. Symp. on Mixed and Augmented Reality, p.121–126.

[29] Ross,D.A., Lim,J., Lin,R.S., , 2008. Incremental learning for robust visual tracking. Int. J. Comput.Vis., 77(1-3):125–141.

[30] Shu,G., Dehghan, A., Oreifej,O. , , 2012. Partbased multiple-person tracking with partial occlusion handling. IEEE Conf. on Computer Vision and Pattern Recognition, p.1815–1821.

[31] Smeulders,A.W., Chu, D., Cucchiara,R. , , 2014. Visual tracking: an experimental survey. IEEE Trans. Patt. Anal. Mach. Intell., 36(7):1442–1468.

[32] Song,S., Xiao,J., 2013. Tracking revisited using RGBD camera: unified benchmark and baselines. IEEE Int. Conf. on Computer Vision, p.233–240.

[33] Teichman,A., Lussier, J.T., Thrun,S. , 2013. Learning to segment and track in RGBD. IEEE Trans. Autom. Sci. Eng., 10(4):841–852.

[34] Wu,Y., Lim,J., Yang,M., 2013. Online object tracking: a benchmark. IEEE Conf. on Computer Vision and Pattern Recognition, p.2411–2418.

[35] Yang,B., Nevatia, R., 2012. Online learned discriminative part-based appearance models for multi-human tracking. ECCV, p.484–498.

[36] Yang,H., Shao,L., Zheng,F., , 2011. Recent advances and trends in visual tracking: a review.Neurocomputing, 74(18):3823–3831.

[37] Yang,M., Wu,Y., Hua,G., 2009. Context-aware visual tracking. IEEE Trans. Patt. Anal. Mach. Intell., 31(7):1195–1209.

[38] Yilmaz,A., Javed, O., Shah,M. , 2006. Object tracking: a survey.ACM Comput. Surv., 38(4):13.

[39] Zhang,L., Maaten, L., 2014. Preserving structure in modelfree tracking. IEEE Trans. Patt. Anal. Mach. Intell., 36(4):756–769.

[40] Zhang,K., Zhang, L., Liu,Q. , , 2014. Fast visual tracking via dense spatio-temporal context learning. ECCV, p.127–141.

RIGHTS & PERMISSIONS

Zhejiang University and Springer-Verlag Berlin Heidelberg