Abstract

We propose a segmentation technique adapted to objects composed of several regions with gray-level fluctuations described by different probability laws. This approach is based on information theory techniques and leads to a multiregion polygonal snake driven by the minimization of a criterion without any parameters to be tuned by the user. We demonstrate the improvements obtained with this approach as well as its low computational cost. This approach is compatible with applications such as object recognition and object tracking with nonrigid deformation in images perturbed by different types of optical noise.

© 2004 Optical Society of America

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. M. Kass, A. Witkin, and D. Terzopoulos, Int. J. Comput. Vision 1, 321 (1988).
    [CrossRef]
  2. R. Ronfard, Int. J. Comput. Vision 13, 229 (1994).
    [CrossRef]
  3. O. Germain and Ph. Réfrégier, Opt. Lett. 21, 1845 (1996).
    [CrossRef] [PubMed]
  4. C. Chesnaud, Ph. Réfrégier, and V. Boulet, IEEE Trans. Pattern Anal. Mach. Intell. 21, 1145 (1999).
    [CrossRef]
  5. F. Galland, N. Bertaux, and Ph. Réfrégier, IEEE Trans. Image Process. 12, 995 (2003).
    [CrossRef]
  6. J. Rissanen, Stochastic Complexity in Statistical Inquiry (World Scientific, Singapore, 1989).
  7. Y. G. Leclerc, Int. J. Comput. Vision 3, 73 (1989).
    [CrossRef]
  8. J. W. Goodman, in Statistical Properties of Laser Speckle Patterns, Vol. 9 of Topics in Applied Physics (Springer-Verlag, Heidelberg, 1975), pp. 9–75.
  9. C. E. Shannon, Bell Syst. Tech. J. 27, 379 (1948).
    [CrossRef]
  10. O. Ruch and Ph. Réfrégier, Opt. Lett. 26, 977 (2001).
    [CrossRef]

2003

F. Galland, N. Bertaux, and Ph. Réfrégier, IEEE Trans. Image Process. 12, 995 (2003).
[CrossRef]

2001

1999

C. Chesnaud, Ph. Réfrégier, and V. Boulet, IEEE Trans. Pattern Anal. Mach. Intell. 21, 1145 (1999).
[CrossRef]

1996

1994

R. Ronfard, Int. J. Comput. Vision 13, 229 (1994).
[CrossRef]

1989

Y. G. Leclerc, Int. J. Comput. Vision 3, 73 (1989).
[CrossRef]

1988

M. Kass, A. Witkin, and D. Terzopoulos, Int. J. Comput. Vision 1, 321 (1988).
[CrossRef]

1948

C. E. Shannon, Bell Syst. Tech. J. 27, 379 (1948).
[CrossRef]

Bertaux, N.

F. Galland, N. Bertaux, and Ph. Réfrégier, IEEE Trans. Image Process. 12, 995 (2003).
[CrossRef]

Boulet, V.

C. Chesnaud, Ph. Réfrégier, and V. Boulet, IEEE Trans. Pattern Anal. Mach. Intell. 21, 1145 (1999).
[CrossRef]

Chesnaud, C.

C. Chesnaud, Ph. Réfrégier, and V. Boulet, IEEE Trans. Pattern Anal. Mach. Intell. 21, 1145 (1999).
[CrossRef]

Galland, F.

F. Galland, N. Bertaux, and Ph. Réfrégier, IEEE Trans. Image Process. 12, 995 (2003).
[CrossRef]

Germain, O.

Goodman, J. W.

J. W. Goodman, in Statistical Properties of Laser Speckle Patterns, Vol. 9 of Topics in Applied Physics (Springer-Verlag, Heidelberg, 1975), pp. 9–75.

Kass, M.

M. Kass, A. Witkin, and D. Terzopoulos, Int. J. Comput. Vision 1, 321 (1988).
[CrossRef]

Leclerc, Y. G.

Y. G. Leclerc, Int. J. Comput. Vision 3, 73 (1989).
[CrossRef]

Réfrégier, Ph.

F. Galland, N. Bertaux, and Ph. Réfrégier, IEEE Trans. Image Process. 12, 995 (2003).
[CrossRef]

O. Ruch and Ph. Réfrégier, Opt. Lett. 26, 977 (2001).
[CrossRef]

C. Chesnaud, Ph. Réfrégier, and V. Boulet, IEEE Trans. Pattern Anal. Mach. Intell. 21, 1145 (1999).
[CrossRef]

O. Germain and Ph. Réfrégier, Opt. Lett. 21, 1845 (1996).
[CrossRef] [PubMed]

Rissanen, J.

J. Rissanen, Stochastic Complexity in Statistical Inquiry (World Scientific, Singapore, 1989).

Ronfard, R.

R. Ronfard, Int. J. Comput. Vision 13, 229 (1994).
[CrossRef]

Ruch, O.

Shannon, C. E.

C. E. Shannon, Bell Syst. Tech. J. 27, 379 (1948).
[CrossRef]

Terzopoulos, D.

M. Kass, A. Witkin, and D. Terzopoulos, Int. J. Comput. Vision 1, 321 (1988).
[CrossRef]

Witkin, A.

M. Kass, A. Witkin, and D. Terzopoulos, Int. J. Comput. Vision 1, 321 (1988).
[CrossRef]

Bell Syst. Tech. J.

C. E. Shannon, Bell Syst. Tech. J. 27, 379 (1948).
[CrossRef]

IEEE Trans. Pattern Anal. Mach. Intell.

C. Chesnaud, Ph. Réfrégier, and V. Boulet, IEEE Trans. Pattern Anal. Mach. Intell. 21, 1145 (1999).
[CrossRef]

IEEE Trans. Image Process.

F. Galland, N. Bertaux, and Ph. Réfrégier, IEEE Trans. Image Process. 12, 995 (2003).
[CrossRef]

Int. J. Comput. Vision

M. Kass, A. Witkin, and D. Terzopoulos, Int. J. Comput. Vision 1, 321 (1988).
[CrossRef]

Int. J. Comput. Vision

R. Ronfard, Int. J. Comput. Vision 13, 229 (1994).
[CrossRef]

Y. G. Leclerc, Int. J. Comput. Vision 3, 73 (1989).
[CrossRef]

Opt. Lett.

Topics in Applied Physics

J. W. Goodman, in Statistical Properties of Laser Speckle Patterns, Vol. 9 of Topics in Applied Physics (Springer-Verlag, Heidelberg, 1975), pp. 9–75.

Other

J. Rissanen, Stochastic Complexity in Statistical Inquiry (World Scientific, Singapore, 1989).

Cited By

OSA participates in CrossRef's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (4)

Fig. 1
Fig. 1

Segmentation results: a, metal brush and its shadow; b, same but corrupted with Gaussian noise (σ=30); c, finger corrupted with Poisson noise; d, a pair of scissors. The image sizes and computation times in milliseconds are written under each image, and the initial and final snakes are shown. Images a, b, and d have been segmented with a Gaussian criterion, and image c with a Poisson criterion.

Fig. 2
Fig. 2

Segmentation and tracking of a truck lowering its dumper (only extracts of whole images are shown). Each image contains 384×288 pixels and is segmented with a Gaussian criterion in 200 ms.

Fig. 3
Fig. 3

Segmentation of a synthetic image (380×266 pixels) corrupted with a gamma pdf of the order of 1 for the car and background regions and a Gaussian pdf for the shadow. d, Segmentation result with the true pdf family (i.e., Gaussian for the shadow and gamma of the order of 1 for the two other regions). e, Segmentation result when the criterion has been determined with three Gaussian pdfs. f, Same as e but with three gamma pdfs of the order of 1. The noisy image gray levels have been modified to be better visualized.

Fig. 4
Fig. 4

Segmentation of a synthetic image (293×271 pixels) corrupted with a geometric pdf for the three center regions and a Poisson pdf for the two other regions. The mean value in each region is written in b. d, Segmentation result with the true pdf family (i.e., geometric for the three center regions and Poisson for the two others). e, Segmentation result when the criterion has been determined with five Poisson pdfs. f, Same as e but with five geometric pdfs. The noisy image gray levels have been modified to be better visualized.

Tables (1)

Tables Icon

Table 1 Expressions of Δr for Different pdfsa

Metrics