We propose a method for tracking moving objects in a nonstationary environment.
While for stationary environment, background substraction has been widely used for object tracking,the same method cannot be applied when the background contains moving components, like watersurface, moving clouds or waving tree branches. Our method combines both of Gaussian Mixture Model and Hidden Markov Model to find the foreground objects. In the training process, Gaussian Mixture Model is used in the first place to obtain an approximated state estimation. Afterwards, the approximated values are fed into HMM as the initial seeds for EM iteration. By that, we hope to overcome the problem of local maxima, usually seen in the training of complicated Hidden Markov Model.
The method is done in a pixelwise manner, followed by a spatial smoothing technique which can resist noise. In the test part, the built pixelwise machines are adopted for real-time object tracking.

We propose a method for tracking moving objects in a nonstationary environment.
While for stationary environment, background substraction has been widely used for object tracking,the same method cannot be applied when the background contains moving components, like watersurface, moving clouds or waving tree branches. Our method combines both of Gaussian Mixture Model and Hidden Markov Model to find the foreground objects. In the training process, Gaussian Mixture Model is used in the first place to obtain an approximated state estimation. Afterwards, the approximated values are fed into HMM as the initial seeds for EM iteration. By that, we hope to overcome the problem of local maxima, usually seen in the training of complicated Hidden Markov Model.
The method is done in a pixelwise manner, followed by a spatial smoothing technique which can resist noise. In the test part, the built pixelwise machines are adopted for real-time object tracking.

Bibliography
[1] A.Iketani, A.Nagai, Y. Kuno, and Y. Shirai. Detecting persons on changing
background. In Proceedings of International Conference on Pattern
Recognition, pages 74–76, 1998.
[2] J.A. Bilmes. A gentle tutorial of em algorithm and its application to
parameter estimation for gaussian mixture and hidden markov models.
Technical Report, University of Berkeley, ICSI-TR-97-021, 1997.
[3] C.Biernacki, G. Celeux, and G. Govaert. Assessing a mixture model for
clustering with the integrated completed likelihood. IEEE Transactions
on Pattern Analysis and Machine Intelligence, 22(7):719–725, 2000.
[4] A.P. Dempster, N.M. Laird, and D.B Rubin. Maximum likelihood from
incomplete data via the em algorithm. J.Roy. Stat. Soc., 39(1):1–38,
1977.
[5] R. Durbin, S. Eddy, A. Krogh, and G. Mitchison. Biological sequence
analysis. The press syndicate of the university of Cambridge, 1998.
[6] M. Figueiredo and A.K. Jain. Unsupervised learning of finite mixture
models. IEEE Transactions on Pattern Analysis and Machine Intelligence,
24(3):381–396, 2002.
[7] W. M. Fitch and T. F. Smith. Optimal sequences alignments. Proceedings
of the National Academy of Science, 80:1382–1386, 1983.
[8] I. Haritaoglu, D. Harwood, and L.S. Davis. w4: real-time surveillance of
people and their activities. IEEE Trans. Pattern Anal. Machine Intell,
22:809–830, 2000.
[9] H. Hartley. Maximum likelihood estimation from incomplete data. Biometrics,
14:174–194, 1958.
[10] Barron J, Fleet D, and Beauchemin S. Performance of optical flow
techniques. Int J Comput Vision, 12(1):42–77, 1994.
[11] M. I. Jordan. Learning in Graphical Models. Kluwer Academic Publishers,
1998.
[12] A.J. Lipton, H. Fujiyoshi, and R.S. Patil. Moving target classification
and tracking from real-time video. in Proc. IEEE Workshop Applications
of Computer Vision, pages 8–14, 1998.
[13] S. McKenna, S. Jabri, Z. Duric, A. Tosenfeld, and H.Wechsler. Tracking
groups of people. Comput. Vis. ImageUnderstanding, 80(1):42–56, 2000.
[14] S.J. Mckenna, S. Gong, and Y. Raja. Tracking colour objects using
adaptive mixture models. Image and Vision Computing, 17(3-4):225–
231, 1999.
[15] D. Meyer, J. Denzler, and H. Niemann. Model based extrection of articulated
objects in image sequences for gait analysis. IEEE International
Conference on Image Processing, Santa Barbara, California, pages 78–
81, 1997.
[16] Thomas P. Minka. The expectation-maximization algorithm for map
estimation. 1998.
[17] S. Roberts, D. Husmeier, I. Rezek, and W. Penny. Bayesian approaches
to gaussian mixture modeling. IEEE Transactions on Pattern Analysis
and Machine Intelligence, 20(11):1133–1142, 1998.
[18] Gideon Schwarz. Estimating the dimension of a model. The Annals of
Statistics, 6(2):461–464, 1978.
[19] C. Stauffer and W. Grimson. Adaptive background mixture models for
real-time tracking. in Proc. IEEE Conf. Computer Vision and Pattern
Recognition, 2:246–252, 1999.
[20] C. Stauffer and W. Grimson. Learning patterns of activity using realtime
tracking. IEEE Trans. Pattern Analysis and Machine Intelligence,
22(8):747–757, 2000.
[21] Carlo Tomasi. Estimating gaussian mixture densities with em - a tutorial.
2003.
[22] K. Toyama, J. Krumm, B. Brumitt, and B. Meyers. Wallflower: Principles
and practice of background maintenance. In Proceedings of IEEE
Int’l Conf. on Computer Vision, pages 255–261, 1999.
[23] A. Verri, S. Uras, and E. DeMicheli. Motion segmentation from optical
flow. Proc the 5th Alvey Vision conference, Brighton, UK, pages 209–
214, 1989.
[24] M. Vingron and M. S. Waterman. Sequence alignment and penalty
choice: Review of concepts, case studies and implications. Journal of
Molecular Biology, 235:1–12, 1994.
[25] L. Wixson. Detecting salient motion by accumulating directionaryconsistenct
flow. IEEE Trans. Pattern Analysis and Machine Intelligence,
22(8):774–780, 2000.
[26] T. Xiang and S. Gong. Beyond tracking: Modeling activity and understanding
behaviour. International Journal of Computer Vision,
67(1):21–51, 2006.
[27] J. Yamato, J. Ohya, and K. Ishii. Recognizing human action in timesequential
images using hidden markov model. Proc. IEEE Conference
on Computer Vision and Pattern Recognition, pages 379–385, 1992.