An Improved Adaptive Kernel-Based Object Tracking

Abstract:

Kernel-based density estimation technique, especially Mean-shift based tracking technique, is a successful application to target tracking, which has the characteristics such as with few parameters, robustness, and fast convergence. However, classic Mean-shift based tracking algorithm uses fixed kernel-bandwidth, which limits the performance when the target’s orientation and scale change. An Improved adaptive kernel-based object tracking is proposed, which extend 2-dimentional mean shift to 3-dimentional, meanwhile combine multiple scale theory into tracking algorithm. Such improvements can enable the algorithm not only track zooming objects, but also track rotating objects. The experimental results validate that the new algorithm can adapt to the changes of orientation and scale of the target effectively.

• Periodical: Advanced Materials Research (Volumes 383-390)
• Pages: 7588-7594
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.383-390.7588
• Online since: November 2011
• Authors: Zheng Hua Liu, Li Han
• Keywords: Kernel Function, Mean-Shift, Orientation, Scale, Target Tracking

References

[1] Weiming Hu, Teiniu Tan, Liang Wang, S. Maybank, A Survey on Visual Surveillance of Object Motion and Behaviors, IEEE Transactions on Systems, Man and Cybernetics, vol. 34(3), Aug. 2004, pp.334-352.

DOI: 10.1109/tsmcc.2004.829274

[2] Yizong Cheng, Mean shift, mode seeking and clustering, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 17(8), Aug. 1995, pp.790-799.

DOI: 10.1109/34.400568

[3] D. Comaniciu, P. Meer, Kernel-based object tracking, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25(3), May 2003, pp.564-575.

DOI: 10.1109/tpami.2003.1195991

[4] R. T. Collins, Mean-shift blob tracking through scale space, Proc. IEEE Conference on Computer Vision and Pattern Recognition, IEEE Press, June 2003, pp.234-240.

DOI: 10.1109/cvpr.2003.1211475

[5] Elgammal, R. Duraiswami, and L. S. Davis, Probabilistic tracking in joint feature-spatial spaces, Proc. IEEE Conference on Computer Vision and Pattern Recognition, IEEE Press, June 2003, p.781–788.

DOI: 10.1109/cvpr.2003.1211432

[6] Changjiang Yang, R. Duraiswami, and L. Davis, Efficient mean-shift via a new similarity measure, Proc. IEEE Conference on Computer Vision and Pattern Recognition, IEEE Press, June 2005, p.176–183.

DOI: 10.1109/cvpr.2005.139

[7] G. D. Hager, M. Dewan, and C. V. Stewart, Multiple kernel tracking with SSD, Proc. IEEE Conference on Computer Vision and Pattern Recognition, IEEE Press, June 2004, p.790–797.

DOI: 10.1109/cvpr.2004.1315112

[8] Z. Zivkovic, B. Krose, An EM-like algorithm for color-histogram-based object tracking, Proc. IEEE Conference on Computer Vision and Pattern Recognition, IEEE Press, June 2004, p.798–803.

DOI: 10.1109/cvpr.2004.1315113

[9] N. Petrovic, L. Jovanov, A. Pizurica, and W. Philips, Efficient video segmentation using temporally updated mean shift clustering, Proc. Security Professionals Information Exchange, IEEE Press, Oct. 2008, pp. 70731R-70731R-10.

DOI: 10.1117/12.792998