Learning and Inferring Motion Patterns using Parametric Segmental Switching Linear Dynamic Systems

  • Sang Min Oh
  • James M. Rehg
  • Tucker Balch
  • Frank Dellaert
Article

Abstract

Switching Linear Dynamic System (SLDS) models are a popular technique for modeling complex nonlinear dynamic systems. An SLDS can describe complex temporal patterns more concisely and accurately than an HMM by using continuous hidden states. However, the use of SLDS models in practical applications is challenging for three reasons. First, exact inference in SLDS models is computationally intractable. Second, the geometric duration model induced in standard SLDSs limits their representational power. Third, standard SLDSs do not provide a principled way to interpret systematic variations governed by higher order parameters.

The contributions in this paper address all of these three challenges. First, we present a data-driven MCMC (DD-MCMC) sampling method for approximate inference in SLDSs. We show DD-MCMC provides an efficient method for estimation and learning in SLDS models. Second, we present segmental SLDSs (S-SLDS), where the geometric distributions of the switching state durations are replaced with arbitrary duration models. Third, we extend the standard SLDS model with additional global parameters that can capture systematic temporal and spatial variations. The resulting parametric SLDS model (P-SLDS) uses EM to robustly interpret parametrized motions by incorporating additional global parameters that underly systematic variations of the overall motion.

The overall development of the extensions for SLDSs provide a principled framework to interpret complex motions. The framework is applied to the honey bee dance interpretation task in the context of the on-going BioTracking project at the Georgia Institute of Technology. The experimental results suggest that the enhanced models provide an effective framework for a wide range of motion analysis applications.

Keywords

Probabilistic graphical models Time-series Trajectory analysis Behavior recognition MCMC Biology 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Andrieu, C., de Freitas, N., Doucet, A., & Jordan, M. I. (2003). An introduction to MCMC for machine learning. Machine Learning, 50, 5–43. MATHCrossRefGoogle Scholar
  2. Balch, T., Dellaert, F., Feldman, A., Guillory, A., Isbell, C., Khan, Z., Stein, A., & Wilde, H. (2006). How A.I. and multi-robot systems research will accelerate our understanding of social animal behavior. Proceedings of IEEE, 94(7), 1145–1463. CrossRefGoogle Scholar
  3. Balch, T., Khan, Z., & Veloso, M. (2001). Automatically tracking and analyzing the behavior of live insect colonies. In Proceedings autonomous agents (pp. 521–528), Montreal. Google Scholar
  4. Bar-Shalom, Y., & Fortmann, T. E. (1988). Tracking and data association. New York: Academic Press. MATHGoogle Scholar
  5. Bar-Shalom, Y., & Li, X. (1993). Estimation and tracking: principles, techniques and software. Boston: Artech House. MATHGoogle Scholar
  6. Bar-Shalom, Y., & Tse, E. (1975). Tracking in a cluttered environment with probabilistic data-association. Automatica, 11, 451–460. MATHCrossRefGoogle Scholar
  7. Barbu, A., & Zhu, S.-C. (2005). Generalizing Swendsen–Wang to sampling arbitrary posterior probabilities. IEEE Transactions on Pattern Analysis and Machine Intelligence, 27(8), 1239–1253. CrossRefGoogle Scholar
  8. Brand, M., & Hertzmann, A. (2000). Style machines. In SIGGRAPH: proceedings of the of conference on computer graphics and interactive technologies (pp. 183–192). Google Scholar
  9. Branson, K., & Belongie, S. (2005). Tracking multiple mouse contours (without too many samples). In Proceedings of IEEE conference on computer vision and pattern recognition (CVPR) (Vol. 1, pp. 1039–1046). Google Scholar
  10. Bregler, C. (1997). Learning and recognizing human dynamics in video sequences. In Proceedings of IEEE conference on computer vision and pattern recognition (CVPR) (pp. 568–574). Google Scholar
  11. Carter, C., & Kohn, R. (1996). Markov chain Monte Carlo in conditionally Gaussian state spaece models. Biometrika, 83, 589–601. MATHCrossRefMathSciNetGoogle Scholar
  12. Casella, G., & Robert, C. P. (1996). Rao-Blackwellisation of sampling schemes. Biometrika, 83(1), 81–94. MATHCrossRefMathSciNetGoogle Scholar
  13. Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society, Series B, 39(1), 1–38. MATHMathSciNetGoogle Scholar
  14. Djuric, P. M., & Chun, J.-H. (2002). An MCMC sampling approach to estimation of nonstationary hidden Markov Models. IEEE Transactions on Signal Processing, 50(5), 1113–1123. CrossRefGoogle Scholar
  15. Dollár S, P., Belongie, S., Branson, K., & Rabaud, V. (2005). Monitoring animal behavior in the smart vivarium. In Measuring behavior (pp. 70–72). Google Scholar
  16. Doucet, A., & Andrieu, C. (2001). Iterative algorithms for state estimation of jump Markov linear systems. IEEE Transactions on Signal Processing, 49(6), 1216–1227. CrossRefGoogle Scholar
  17. Doucet, A., Gordon, N. J., & Krishnamurthy, V. (2001). Particle filters for state estimation of jump Markov linear systems. IEEE Transactions on Signal Processing, 49(3), 613–624. CrossRefGoogle Scholar
  18. Feldman, A., & Balch, T. (2004). Representing honey bee behavior for recognition using human trainable models. Adaptive Behavior, 12, 241–250. CrossRefGoogle Scholar
  19. Ferguson, J. (1980). Variable duration models for speech. In Symposium on the application of HMMs to text and speech (pp. 143–179). Google Scholar
  20. Frey, B., & Jojic, N. (2003). Transformation-invariant clustering and dimensionality reduction using EM. IEEE Transactions on Pattern Analysis and Machine Intelligence, 25(1), 1–17. CrossRefGoogle Scholar
  21. Frisch, K. (1967). The dance language and orientation of bees. Harvard: Harvard University Press. Google Scholar
  22. Ge, X., & Smyth, P. (2000). Deformable Markov model templates for time-series pattern matching. In International conference knowledge discovery and data mining (KDD) (pp. 81–90). Google Scholar
  23. Ghahramani, Z., & Hinton, G. E. (1998). Variational learning for switching state-space models. Neural Computation, 12(4), 963–996. Google Scholar
  24. Gilks, W. R., Richardson, S., & Spiegelhalter, D. J. (Eds). (1996). Markov chain Monte Carlo in practice. London: Chapman and Hall. MATHGoogle Scholar
  25. Hastings, W. K. (1970). Monte Carlo sampling methods using Markov chains and their applications. Biometrika, 57, 97–109. MATHCrossRefGoogle Scholar
  26. Howard, A., & Jebara, T. (2004). Dynamical systems trees. In Proceedings of the 20th conference on uncertainty in AI (UAI) (pp. 260–267), Banff, Canada. Google Scholar
  27. Khan, Z., Balch, T., & Dellaert, F. (2004). A Rao-Blackwellized particle filter for eigentracking. In Proceedings of IEEE conference on computer vision and pattern Recognition (CVPR) (Vol. 2, pp. 980–986). Google Scholar
  28. Khan, Z., Balch, T., & Dellaert, F. (2006). MCMC data association and sparse factorization updating for real time multitarget tracking with merged and multiple measurements. IEEE Transactions on Pattern Analysis and Machine Intelligence, 28(12), 1960–1972. CrossRefGoogle Scholar
  29. Kim, C.-J. (1994). Dynamic linear models with Markov-switching. Journal of Econometrics, 60(1–2), 1–22. MATHCrossRefMathSciNetGoogle Scholar
  30. Kim, S., & Smyth, P. (2006). Segmental Hidden Markov Models with Random Effects for Waveform Modeling. Journal of Machine Learning Research, 7, 945–969. MathSciNetGoogle Scholar
  31. Lee, M. W., & Cohen, I. (2006). A model-based approach for estimating human 3d poses in static images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 28(6), 905–916. CrossRefGoogle Scholar
  32. Lerner, U., & Parr, R. (2001). Inference in hybrid networks: theoretical limits and practical algorithms. In Proceedings of the 17th conference on uncertainty in AI (UAI) (pp. 310–318), Seattle, WA, August 2001. Google Scholar
  33. Lerner, U., Parr, R., Koller, D., & Biswas, G. (2000). Bayesian fault detection and diagnosis in dynamic systems. In Proceedings of the 17th AAAI national conference on AI (pp. 531–537), Austin, TX. Google Scholar
  34. Levinson, S. E. (1990). Continuously variable duration hidden Markov models for automatic speech recognition. Computer Speech and Language, 1(1), 29–45. CrossRefGoogle Scholar
  35. Li, Y., Wang, T., & Shum, H-Y. (2002). Motion texture: a two-level statistical model for character motion synthesis. In SIGGRAPH: proceedings of the conference on computer graphics and interactive technologies. Google Scholar
  36. Maybeck, P. (1979). Stochastic models, estimation and control (Vol. 1). New York: Academic Press. MATHGoogle Scholar
  37. McLachlan, G. J., & Krishnan, T. (1997). Wiley series in probability and statistics. The EM algorithm and extensions. New York: Wiley. MATHGoogle Scholar
  38. Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., & Teller, E. (1953). Equations of state calculations by fast computing machine. Journal of Chemical Physics, 21, 1087–1091. CrossRefGoogle Scholar
  39. Neal, R. M., & Hinton, G. E. (1998). A view of the EM algorithm that justifies incremental, sparse, and other variants. Dordrecht: Kluwer Academic. Also published by MIT Press, 1999. Google Scholar
  40. North, B., Blake, A., Isard, M., & Rottscher, J. (2000). Learning and classification of complex dynamics. IEEE Transactions on Pattern Analysis and Machine Intelligence, 22(9), 1016–1034. CrossRefGoogle Scholar
  41. Oh, S. M., Ranganathan, A., Rehg, J. M., & Dellaert, F. (2005a). A variational inference method for switching linear dynamic systems. GVU Center, College of Computing: Technical Report GIT-GVU-05-16. Google Scholar
  42. Oh, S. M., Rehg, J. M., Balch, T., & Dellaert, F. (2005b). Data-driven MCMC for learning and inference in switching linear dynamic systems. In Proceedings of the 22nd AAAI national conference on AI (pp. 944–949), Pittsburgh, PA. Google Scholar
  43. Oh, S. M., Rehg, J. M., Balch, T., & Dellaert, F. (2005c). Learning and inference in parametric switching linear dynamic systems. In Proceedings of the international conference on computer vision (ICCV) (Vol. 2, pp. 1161–1168). Google Scholar
  44. Ostendorf, M., Digalakis, V. V., & Kimball, O. A. (1996). From HMM’s to segment models: a unified view of stochastic modeling for speech recognition. IEEE Transactions on Speech and Audio Processing, 4(5), 360–378. CrossRefGoogle Scholar
  45. Pavlović, V., & Rehg, J. M. (2000). Impact of dynamic model learning on classification of human motion. In Proceedings of IEEE conference on computer vision and pattern recognition (CVPR) (Vol. 1, pp. 788–795). Google Scholar
  46. Pavlović, V., Rehg, J. M., Cham, T.-J., & Murphy, K. (1999). A dynamic Bayesian network approach to figure tracking using learned dynamic models. In Proceedings of the international conference on computer vision (ICCV) (Vol. 1, pp. 94–101). Google Scholar
  47. Pavlović, V., Rehg, J. M., & MacCormick, J. (2000) Learning switching linear models of human motion. In Advances in neural information processing systems (NIPS) (pp. 981–987). Google Scholar
  48. Ranganathan, A., & Dellaert, F. (2005). Data driven MCMC for appearance-based topological mapping. In Robotics: science and systems I (pp. 209–216). Google Scholar
  49. Ren, L., Patrick, A., Efros, A., Hodgins, J., & Rehg, J. M. (2005). A data-driven approach to quantifying natural human motion. ACM Transactions on Graphics, Special Issue: Proceedings of 2005 SIGGRAPH Conference, 24(3), 1090–1097. Google Scholar
  50. Rosti, A.-V. I., & Gales, M. J. F. (2004). Rao-Blackwellised Gibbs sampling for switching linear dynamical systems. In Proceedings of international conference on acoustic, speech, and signal processing (ICASSP) (Vol. 1, pp. 809–812). Google Scholar
  51. Roweis, S., & Ghahramani, Z. (1999). A unifying review of linear Gaussian models. Neural Computation, 11(2), 305–345. CrossRefGoogle Scholar
  52. Russel, M. (1993). A segmental HMM for speech pattern matching. In Proceedings of international conference on acoustic, speech, and signal processing (ICASSP) (pp. 499–502). Google Scholar
  53. Shumway, R. H., & Stoffer, D. S. (1992). Dynamic linear models with switching. Journal of the American Statistical Association, 86, 763–769. CrossRefMathSciNetGoogle Scholar
  54. Tu, Z. W., & Zhu, S. C. (2002). Image segmentation by data-driven Markov chain Monte Carlo. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24(5), 657–673. CrossRefGoogle Scholar
  55. Vidal, R., Chiuso, A., & Soatto, S. (2002). Observability and identifiability of jump linear systems. In Proceedings of IEEE conference on decision and control (Vol. 4, pp. 3614–3619). Google Scholar
  56. Wilson, A. D., & Bobick, A. F. (1999). Parametric hidden Markov models for gesture recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 21(9), 884–900. CrossRefGoogle Scholar
  57. Zoeter, O., & Heskes, T. (2003). Hierarchical visualization of time-series data using switching linear dynamical systems. IEEE Transactions on Pattern Analysis and Machine Intelligence, 25(10), 1202–1215. CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2007

Authors and Affiliations

  • Sang Min Oh
    • 1
  • James M. Rehg
    • 1
  • Tucker Balch
    • 1
  • Frank Dellaert
    • 1
  1. 1.GVU Center, College of ComputingGeorgia Institute of TechnologyAtlantaUSA

Personalised recommendations