Variational Segmentation of Image Sequences Using Deformable Shape Priors

  • Ketut Fundana
  • Niels Chr. Overgaard
  • Anders Heyden
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4522)


The segmentation of objects in image sequences is an important and difficult problem in computer vision with applications to e.g. video surveillance. In this paper we propose a new method for variational segmentation of image sequences containing nonrigid, moving objects. The method is based on the classical Chan-Vese model augmented with a novel frame-to-frame interaction term, which allow us to update the segmentation result from one image frame to the next using the previous segmentation result as a shape prior. The interaction term is constructed to be pose-invariant and to allow moderate deformations in shape. It is expected to handle the appearance of occlusions which otherwise can make segmentation fail. The performance of the model is illustrated with experiments on real image sequences.


Variational formulation segmentation tracking region-based level sets interaction terms deformable shape priors 


  1. 1.
    Caselles, V., Kimmel, R., Sapiro, G.: Geodesic active contours. International Journal of Computer Vision 22(1), 61–79 (1997)zbMATHCrossRefGoogle Scholar
  2. 2.
    Chan, T., Vese, L.: Active contour without edges. IEEE Transactions on Image Processing 10(2), 266–277 (2001)zbMATHCrossRefGoogle Scholar
  3. 3.
    Chan, T., Zhu, W.: Level set based prior segmentation. Technical Report UCLA CAM 03-66, Department of Mathematics, UCLA (2003)Google Scholar
  4. 4.
    Cremers, D.: Statistical Shape Knowledge in Variational Image Segmentation. Phd thesis, Department of Mathematics and Computer Science, University of Mannheim (July 2002)Google Scholar
  5. 5.
    Cremers, D., Sochen, N.: Towards recognition-based variational segmentation using shape priors and dynamic labeling. In: Griffin, L.D., Lillholm, M. (eds.) Scale Space Methods in Computer Vision. LNCS, vol. 2695, pp. 388–400. Springer, Heidelberg (2003)CrossRefGoogle Scholar
  6. 6.
    Cremers, D., Soatto, S.: A pseudo-distance for shape priors in level set segmentation. In: Faugeras, O., Paragios, N. (eds.) 2nd IEEE Workshop on Variational, Geometric and Level Set Methods in Computer Vision, IEEE Computer Society Press, Los Alamitos (2003)Google Scholar
  7. 7.
    Cremers, D., Funka-Lea, G.: Dynamical Statistical Shape Priors for Level Set Based Sequence Segmentation. In: Paragios, N., Faugeras, O., Chan, T., Schnörr, C. (eds.) VLSM 2005. LNCS, vol. 3752, pp. 210–221. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  8. 8.
    Rousson, M., Paragios, N.: Shape priors for level set representations. In: Heyden, A., Sparr, G., Nielsen, M., Johansen, P. (eds.) ECCV 2002. LNCS, vol. 2351, pp. 78–92. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  9. 9.
    Leventon, M.E., Grimson, W.E.L., Faugeras, O.: Statistical shape influence in geodesic active contours. In: CVPR (2000)Google Scholar
  10. 10.
    Moelich, M., Chan, T.: Tracking objects with the chan-vese algorithm. Technical Report UCLA CAM 03-14, Department of Mathematics, UCLA (March 2003)Google Scholar
  11. 11.
    Paragios, N., Deriche, R.: Geodesic active contours and level set methods for the detection and tracking of moving objects. IEEE Trans. PAMI 22(3), 266–280 (2000)Google Scholar
  12. 12.
    Paragios, N., Deriche, R.: Geodesic active regions and level set methods for motion estimation and tracking. Computer Vision and Image Understanding 97, 259–282 (2005)CrossRefGoogle Scholar
  13. 13.
    Osher, S., Fedkiw, R.: Level Set Methods and Dynamic Implicit Surfaces. Springer, New York (2003)zbMATHGoogle Scholar
  14. 14.
    Solem, J.E., Overgaard, N.C.: A Geometric Formulation of Gradient Descent for Variational Problems with Moving Surfaces. In: Kimmel, R., Sochen, N.A., Weickert, J. (eds.) Scale-Space 2005. LNCS, vol. 3459, pp. 419–430. Springer, Heidelberg (2005)Google Scholar

Copyright information

© Springer Berlin Heidelberg 2007

Authors and Affiliations

  • Ketut Fundana
    • 1
  • Niels Chr. Overgaard
    • 1
  • Anders Heyden
    • 1
  1. 1.Applied Mathematics Group, School of Technology and Society, Malmö UniversitySweden

Personalised recommendations