Deformable Object Tracking: A Kernel Density Estimation Approach Via Level Set Function Evolution

  • Nilanjan Ray
  • Baidya Nath Saha
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4815)

Abstract

Automated tracking of deformable objects that change shape and size drastically is challenging. For useful results, one needs an efficient deformable object model. In this regard, we propose a novel deformable object model via joint probability density of level set function and image intensity/feature values. Given the delineated object boundary on the first image frame of a video sequence, we learn the aforementioned joint probability density via kernel (Parzen window) method. From the next frame onward, we match this learned probability density with the probability density on the current frame by minimizing Kullback-Leibler divergence. This minimization procedure is cast in a variational framework and a minimizer is obtained by solving a partial differential equation (PDE). A stable and efficient numerical scheme is proposed for solving this resulting PDE. We demonstrate the efficacy of the proposed tracking method on myocardial border tracking from mouse heart cine magnetic resonance imagery (MRI).

Keywords

Kernel density estimation Parzen window KL divergence level set cine MRI 

References

  1. 1.
    Janiczek, R., Ray, N., Acton, S.T., Roy, R.J., French, B.A., Epstein, F.H.: Markov chain Monte Carlo method for tracking myocardial borders. Computational Imaging III. In: Bouman, C.A., Miller, E.L. (eds.) Proceedings of the SPIE, vol. 5674, pp. 211–218 (2005)Google Scholar
  2. 2.
    Comaniciu, D., Ramesh, V., Meer, P.: Kernel-Based Object Tracking. IEEE PAMI 25(5), 564–577 (2003)Google Scholar
  3. 3.
    Montagnat, T., Delingette, H., Ayache, N.: A review of deformable surfaces: topology, geometry and deformation. Image and Vision Computing 19(14), 1023–1040 (2001)CrossRefGoogle Scholar
  4. 4.
    Lai, C., Chin, R.: Deformable Contours: Modeling and Extraction. IEEE PAMI 17(11), 1084–1090 (1995)Google Scholar
  5. 5.
    Cootes, T.F., Edwards, G.J., Taylor, C.J.: Active appearance models. IEEE PAMI 23(6), 681–685 (2001)Google Scholar
  6. 6.
    Kass, M., Witkin, A., Terzopoulos, D.: Snakes: Active contour models. International J. of Comp. Vis. 1(4), 1405–1573 (1988)Google Scholar
  7. 7.
    Sethian, J.A.: Level Set Methods and Fast Marching Methods: Evolving Interfaces in Computational Geometry, Fluid Mechanics, Computer Vision and Materials Sciences. Cambridge University Press, Cambridge (1999)Google Scholar
  8. 8.
    Leventon, M.: Statistical Models for Medical Image Analysis. Ph.D. Thesis, MIT (2000)Google Scholar
  9. 9.
    Freedman, D., Zhang, T.: Active contours for tracking distributions. IEEE Trans. Image Proc. 13(4), 518–526 (2004)CrossRefGoogle Scholar
  10. 10.
    Bishop, C.M.: Pattern Recognition and Machine Learning. Springer, Heidelberg (2006)MATHGoogle Scholar
  11. 11.
    Ames, W.F.: Numerical methods for partial differential equations. Academic, New York (1992)MATHGoogle Scholar
  12. 12.
    Xu, C., Prince, J.L.: Snakes, Shapes, and Gradient Vector Flow. IEEE Transactions on Image Processing 7(3), 359–369 (1998)MATHCrossRefMathSciNetGoogle Scholar
  13. 13.
    Pratt, W.K.: Digital Image Processing. Wiley InterScience, Chichester (2002)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Nilanjan Ray
    • 1
  • Baidya Nath Saha
    • 1
  1. 1.Department of Computing Science, University of Alberta, EdmontonCanada

Personalised recommendations