Classification and Quantification of Occlusion Using Hidden Markov Model

  • C. R. Sahoo
  • Shamik Sural
  • Gerhard Rigoll
  • A. Sanchez
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6744)


Over the last few years, gait recognition has become an active area of research. However, one of the shortcomings is lack of a method for quantifying occlusion in scenes used for capturing gait of individuals. Occlusion can occur primarily because of two reasons. Firstly, movement of certain body parts of a human being occludes some other body parts of the same human, which is called self occlusion and secondly, occlusion of the body parts caused by some other human being. The objective of this paper is to quantify occlusion of different parts of the human body using Hidden Markov Model (HMM) and classify the scene of occlusion as one of the three cases of occlusion, namely, self occlusion (single individual moving), occlusion in a crowd moving in same direction and occlusion due to movement of human beings approaching from opposite direction. We train one HMM for each body part relevant for gait recognition. An HMM is a statistical representation of probability distribution of a large number of possible sequences and in the current context these are the sequences of frames extracted at regular interval from a given video. The steps involved in achieving the objective are feature extraction, HMM training and finally the classification or hidden state generation.


Occlusion Gait Recognition Body Part HMM 


  1. 1.
    Apostoloff, N., Fitzgibbon, A.: Learning spatiotemporal t-junctions for occlusion detection. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, June 2005, vol. 2, pp. 553–559 (2005)Google Scholar
  2. 2.
    Marchesotti, L., Piva, S., Regazzoni, C.S.: A dynamic model integrating colour and shape information for objects tracking in conditions of occlusion. In: IEEE International Conference on Multimedia and Expo, June 2004, vol. 3, pp. 1547–1550 (2004)Google Scholar
  3. 3.
    Mohamed, M.A., Gader, P.: Generalized hidden markov models. ii. application to handwritten word recognition. IEEE Transactions on Fuzzy Systems 8(1), 82–94 (2000)CrossRefGoogle Scholar
  4. 4.
    Pan, J., Hu, B.: Robust occlusion handling in object tracking. In: IEEE Conference on Computer Vision and Pattern Recognition, June 2007, pp. 1–8 (2007)Google Scholar
  5. 5.
    Rabiner, L.R.: A tutorial on hidden markov models and selected applications in speech recognition. Proceedings of the IEEE 77(2), 257–286 (1989)CrossRefGoogle Scholar
  6. 6.
    Veeravalli, A.G., Pan, W.D., Adhami, R., Cox, P.G.: A tutorial on using hidden markov models for phoneme recognition. In: Proceedings of the Thirty-Seventh Southeastern Symposium on System Theory, March 2005, pp. 154–157 (2005)Google Scholar
  7. 7.
    Velipasalar, S., Wolf, W.: Multiple object tracking and occlusion handling by information exchange between uncalibrated cameras. In: IEEE International Conference on Image Processing, September 2005, vol. 2, pp. II–418–II–421 (2005)Google Scholar
  8. 8.
    Wang, L., Hu, W., Tan, T.: Recent developments in human motion analysis. Pattern Recognition 36(3), 585–601 (2003)CrossRefGoogle Scholar
  9. 9.
    Wang, Y., Liu, Z., Zhou, L.: Learning hierarchical non-parametric hidden markov model of human motion. In: Proceedings of International Conference on Machine Learning and Cybernetics, August 2005, vol. 6, pp. 3315–3320 (2005)Google Scholar
  10. 10.
    Yilmaz, A., Li, X., Shah, M.: Contour-based object tracking with occlusion handling in video acquired using mobile cameras. IEEE Transactions on Pattern Analysis and Machine Intelligence 26(11), 1531–1536 (2004)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • C. R. Sahoo
    • 1
  • Shamik Sural
    • 1
  • Gerhard Rigoll
    • 2
  • A. Sanchez
    • 3
  1. 1.School of Information TechnologyIndian Institute of TechnologyKharagpurIndia
  2. 2.Technical University of MunichMunichGermany
  3. 3.Universidad Rey Juan CarlosMadridSpain

Personalised recommendations