Abstract
Particle filters are computationally intensive and thus efficient parallelism is crucial to effective implementations, especially object tracking in video sequences. Two schemes for pipelining particles under high performance computing environment, including an alternative Markov Chain Monte Carlo (MCMC) resampling algorithm and kernel function, are proposed so as to improve tracking performance and minimize execution time. Experimental results on a network of workstations composed of simple off-the-shelf hardware components show that global parallelizable scheme provides a promising resolution to clearly reduce execution time with increasing particles, compared with generic particle filtering.
Similar content being viewed by others
References
Bolic M, Djuric PM, Hong S (2004) Resampling algorithms for particle filters: a computational complexity perspective. EURASIP J Applied Signal Process 44(1–2):2267–2277
Chang C, Ansari R (2005) Kernel particle filter for visual tracking. IEEE Signal Process Lett 12(3):242–245
Chib S, Greenberg E (1995) Understanding the Metropolis–Hastings algorithm. J Am Stat 49:327–335 (1995)
Cho JU, Jin SH, Pham XD, Jeon JW (2007) Multiple objects tracking circuit using particle filters with multiple features. In: IEEE int. conf. on robotics and automation, pp 4639-4644. Italy
Doucet A, Godsill S, Andreieu C (2000) On sequential Monte Carlo sampling methods for Bayesian filtering. J Stat Comput 10:197–208
Higuchi T (1997) Monte Carlo filter using the genetic algorithm operators. J Stat Comput Simul 59(1):1–23
Hong S, Chin S, Djuric P (2006) Design and implementation of flexible resampling mechanism for high-speed parallel particle filters. J. VSLI Signal Process 44:47–62
Huang AJ (2005) A tutorial on Bayesian estimation and tracking techniques applicable to nonlinear and non-Gaussian processes. MIRTE Technique report
Kotecha J, Djuric P (2003) Gaussian particle filtering. IEEE Trans Signal Process 51(10):2592–2601
Liu JS, Chen R, Logvinenko T (2001) A theoretical framework for sequential importance sampling and resampling. In: Sequential Monte Carlo methods in practice. Lecture notes in computer science. Springer, New York
Maskell S, Alun B, Macleod M (2006) A single instruction multiple data particle filter. In: Proc. of nonlinear statistical signal processing workshop
Miguez J (2007) Analysis of parallelizable resampling algorithms for particle filtering. IEEE Trans Signal Process 87:3155–3174
Musa ZB, Watada J (2008) Motion tracking using particle filter. Lect Notes Comput Sci (Springer) 5179:119–126.
Perez P, Vermaak V, Blake A (2004) Data fusion for visual tracking with particles. Proc IEEE 92:495–513
Pitt M, Shepard N (1999) Filtering via simulation: auxiliary particle filters. J Am Stat Assoc 94(446):590–599
Quinn MJ (2003) Parallel programming in C with MPI and OpenMP. McGraw Hill
Ristic B, Arulampalam A, Gordon N (2004) Beyond the Kalman filter: particle filters for tracking applications. Artech House, Boston
Rui Y, Chen Y (2001) Better proposal distributions: object tracking using unscented particle filter. In: Proc. of computer vision and patten recognition, pp 786–793. Kauai
Sutharsan S, Sinha A, Kirubarajan T, Farooq M (2005) An optimization based parallel particle filter for multitarget tracking. Proc SPIE 5913:87–98
Zhou SK, Chellappa R, Moghaddam B (2004) Visual tracking and recognition using appearance-adaptive models in particle filters. IEEE Trans Image Process (11):1434–1456
Acknowledgement
This research was supported by the National Science Foundation of China under Grants 60572041 and 60832004.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Wang, D., Zhang, Q. & Morris, J. Distributed Markov Chain Monte Carlo kernel based particle filtering for object tracking. Multimed Tools Appl 56, 303–314 (2012). https://doi.org/10.1007/s11042-010-0646-4
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-010-0646-4