Robust Dual-Kernel Tracking Using Both Foreground and Background - Advances in Image and Graphics Technologies - page 89

Table 1. The mean value of center location errors (in Pixels)

KBT BC-KBT DKT

Sequence 1

138.9 137.0 5.4

Sequence 2

62.5 59.1 17.9

5 Conclusion

In conclusion, we proposed a dual-kernel tracker based mean shift algorithm using
both foreground and background. The target model consists of foreground model and
background model, and the optimizing process integrates foreground kernel iteration
and background kernel iteration. We update the background model each iteration
to adaptive the background change, and relatively retain the foreground model to
weaken the model drift. The proposed tracker is more robust in coping with the back-
ground change and clutters. The subjective tracking results and quantitative evalua-
tion demonstrate that the proposed dual-kernel tracker outperforms the single-kernel
trackers. An effective method to estimate the scale change may greatly improve the
tracking performance and thus, our next work will focus the scale estimation during
the iteration of two kernels.

Acknowledgements. This research was supported by National Natural Science Foun-
dation of China (No. 61175029).

References

1. Comaniciu, D., Ramesh, V., Meer, P.: Kernel-Based Object Tracking. IEEE Transactions
on Pattern Analysis and Machine Intelligence 25(5), 564-577 (2003)
2. Cheng, Y.: Mean shift, mode seeking, and clustering. IEEE Transactions on Pattern Analy-
sis and Machine Intelligence 17(8), 790-799 (1995)
3. Comaniciu, D., Meer, P.: Mean shift: A robust approach toward feature space analysis.
IEEE Transactions on Pattern Analysis and Machine Intelligence 24(5), 603-619 (2002)
4. Collins, R.T.: Mean-shift blob tracking through scale space. In: IEEE Computer Society
Conference on Computer Vision and Pattern Recognition, pp. 234-240. IEEE Press, Wis-
consin (2003)
5. Birchfield, S.T., Rangarajan, S.: Spatiograms versus histograms for region -based tracking.
In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition,
pp. 1158-1163. IEEE Press, Santiago (2005)
6. Li, P.H.: An Adaptive Binning Color Model for Mean Shift Tracking. IEEE Transactions
on Circuits and Systems for Video Technology 18(9), 1293-1299 (2008)
7. Qi, S., Huang, X., Yi, H.: Object tracking by anisotropic kernel mean shift. Journal of
Electronics and Information Technology 29(3), 686-689 (2007)
8. Liu, R., Jing, Z.: Robust kernel-based tracking algorithm with background contrasting.
Chinese Optical Letters 10(2), 021001 (2012)