Science China Information Sciences

, Volume 53, Issue 7, pp 1336–1344 | Cite as

Learning interactions among multi-channel sequences with dynamical influence models

Research Papers

Abstract

Many real applications involve simultaneous recording and analysis of multi-channel information sources. Learning and modeling the interactions among channels is the kernel step to analyze and recognize system characteristics. This paper presents a model that learns the dynamical influence among multi-channel sequences. The model, dynamical influence model, permits functional roles of individual channels to change and models the changing influence strength between channels. By querying the values of influence factors, we can recognize the functional role of each channel qualitatively and learn about to what extent the chains influence each other quantitatively at any time. The experimental results on synthetic data and application of multi-person interaction recognition show that our model is reliable and effective.

Keywords

multi-channel processing dynamical interaction influence model activity recognition 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Bengio S, Bourlard H. Multi-channel sequence processing. In: Deterministic and Statistical Methods in Machine Learning. Sheffield, UK, 2004. 22–36Google Scholar
  2. 2.
    Poh N, Bengio S. Why do multi-stream, multi-band and multi-modal approaches work on biometric user authentication tasks? In: Proc ICASSP, Montreal, QC, Canada, 2004. 893–896Google Scholar
  3. 3.
    Potamianos G, Neti C, Gravier G, et al. Recent advances in the automatic recognition of audio-visual speech. Proc IEEE, 2003, 91: 1306–1326CrossRefGoogle Scholar
  4. 4.
    Brand M, Oliver N, Pentland A. Coupled hidden Markov models for complex action recognition. In: Proc CVPR, San Juan, Puerto Rico, 1997. 994–999Google Scholar
  5. 5.
    Vogler C, Metaxas D. Parallel hidden Markov models for American sign language recognition. In: Proc ICCV, Kerkyra, Corfu, Greece, 1999. 224–228Google Scholar
  6. 6.
    Asavathiratham C, Roy S, Lesieutre B. The influence model. IEEE Control Syst Mag, 2001, 21: 52–64CrossRefGoogle Scholar
  7. 7.
    Basu S, Choudhury T, Clarkson B, et al. Learning human interactions with the influence model. In: Proc NIPS Vancouver, British Columbia, Canada, 2001Google Scholar
  8. 8.
    Choudhury T, Basu S. Modeling conversational dynamics as a mixed-memory Markov process. In: Proc NIPS, Vancouver, Canada, 2004. 281–288Google Scholar
  9. 9.
    Tian Y, Mei Z, Huang T, et al. Incremental learning for interaction dynamics with the influence model. In: Proc. SIGKDD, Washington, DC, USA, 2003Google Scholar
  10. 10.
    Dong W, Pentland A. Modeling influence between experts. In: AI for Human Computing, LNAI, Hyderadab, India, 2007. 170–189Google Scholar
  11. 11.
    Dong W, Lepri B, Cappelletti A, et al. Using the influence model to recognize functional roles in meetings. In: Proc. ICMI, Nagoya, Japan, 2007. 271–278Google Scholar
  12. 12.
    Rienks R, Zhang D, Gatica-Perez D, et al. Detection and application of influence ranking in small group meetings. In: Proc. ICMI, Banff, Alberta, Canada, 2006. 257–264Google Scholar
  13. 13.
    Bradicka O, Maisonnasse J, Reignier P. Automatic detection of influence groups. In: Proc. ICMI, Trento, Italy, 2005Google Scholar
  14. 14.
    Lee M, Ofsche R. The impact of behavioral style and status characteristics on social influence: a test of two competing theories. Social Psych Quart, 1981, 44: 73–82CrossRefGoogle Scholar
  15. 15.
    Odell J, Parunak H V, Brueckner S, et al. Changing roles: dynamic role assignment. J Object Tech, 2003, 2: 77–86Google Scholar
  16. 16.
    Saul L K, Jordan M I. Mixed memory Markov models: decomposing complex stochastic processes as mixtures of simpler ones. Mach Learn, 1999, 37: 75–87MATHCrossRefGoogle Scholar
  17. 17.
    Rachel A M. Mixed hidden Markov models: an extension of the hidden Markov model to the longitudinal data setting. J Am Stat Assoc, 2007, 102: 201–210MATHCrossRefGoogle Scholar
  18. 18.
    Murphy K P. Dynamic Bayesian networks: representation, inference and learning. Phd thesis, University of California at Berkeley, Computer Science Division, 2002Google Scholar
  19. 19.
    Zhang W, Chen F, Xu W, et al. Real-time video intelligent surveillance system. In: Proc. ICME, Toranto, Ontario, Canada, 2006. 1021–1024Google Scholar
  20. 20.
    Zhang W, Chen F, Xu W, et al. Decomposition in hidden Markov models for activity recognition. In: Proc. MCAM, Weihai, Shandong, China, 2007. 232–241Google Scholar
  21. 21.
    Zhang W, Chen F, Xu W. Hierarchical group process representation in multi-agent activity recognition. Signal Process Image Commun, 2008, 23: 739–753CrossRefGoogle Scholar

Copyright information

© Science China Press and Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  1. 1.Department of Automation, National Laboratory for Information Science and TechnologyTsinghua UniversityBeijingChina

Personalised recommendations