Relaxing Symmetric Multiple Windows Stereo Using Markov Random Fields
This paper introduces R-SMW, a new algorithm for stereo matching. The main aspect is the introduction of a Markov Random Field (MRF) model in the Symmetric Multiple Windows (SMW) stereo algorithm in order to obtain a non-deterministic relaxation. The SMW algorithm is an adaptive, multiple window scheme using left-right consistency to compute disparity. The MRF approach allows to combine in a single functional the disparity values coming from different windows, the left-right consistency constraint and regularization hypotheses. The optimal estimate of the disparity is obtained by minimizing an energy functional with simulated annealing. Results with both synthetic and real stereo pairs demonstrate the improvement over the original SMW algorithm, which was already proven to perform better than state-of-the-art algorithms.
KeywordsMarkov Random Field Observation Model Stereo Match Stereo Pair Markov Random
Unable to display preview. Download preview PDF.
- 1.G.S Nadabar A.K. Jain. Range image segmentation using MRF models. In A.K. Jain and R. Chellappa, editors, Markov Random Fields Theory and Application, pages 542–572. Academic Press, 1993.Google Scholar
- 2.I. J. Cox, S. Hingorani, B. M. Maggs, and S. B. Rao. A maximum likelihood stereo algorithm. Computer Vision and Image Understanding, 63(3):542–567, May 1996.Google Scholar
- 4.U. R. Dhond and J. K. Aggarwal. Structure from stereo-a review. IEEE Transactions on Systems, Man and Cybernetics, 19(6):1489–1510, November/December 1989.Google Scholar
- 5.O. Faugeras, B. Hotz, H. Mathieu, T. Viéville, Z. Zhang, P. Fua, E. Théron, L. Moll, G. Berry, J. Vuillemin, P. Bertin, and C. Proy. Real-time correlation-based stereo: algorithm, implementation and applications. Technical Report 2013, Unité de recherche INRIA Sophia-Antipolis, August 1993.Google Scholar
- 6.P. Fua. Combining stereo and monocular information to compute dense depth maps that preserve depth discontinuities. In Proceedings of the International Joint Conference on Artificial Intelligence, pages 1292–1298, Sydney, Australia, August 1991.Google Scholar
- 7.A. Fusiello, V. Roberto, and E. Trucco. Efficient stereo with multiple windowing. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 858–863, Puerto Rico, June 1997. IEEE Computer Society Press.Google Scholar
- 8.A. Fusiello, V. Roberto, and E. Trucco. Symmetric stereo with multiple windowing. International Journal of Pattern Recognition and Artificial Intelligence, 14(8):1053–1066, December 2000.Google Scholar
- 10.E. Gamble and T. Poggio. Visual integration and detection of discontinuities: the key role of intensity edge. A.I. Memo 970, Massachusetts Institute of Technology, 1987.Google Scholar
- 11.D. Geiger, B. Ladendorf, and A. Yuille. Occlusions and binocular stereo. International Journal of Computer Vision, 14(3):211–226, April 1995.Google Scholar
- 13.S.S. Intille and A. F. Bobick. Disparity-space images and large occlusion stereo. In Jan-Olof Eklundh, editor, European Conference on Computer Vision, pages 179–186, Stockholm, Sweden, May 1994. Springer-Verlag.Google Scholar
- 14.J. Konrad and E. Dubois. Bayesian estimation of motion vector field. IEEE Transactions on Pattern Analysis and Machine Intelligence, 14(9):910–927, September 1992.Google Scholar
- 15.T. Kanade and M. Okutomi. A stereo matching algorithm with an adaptive window: Theory and experiments. IEEE Transactions on Pattern Analysis and Machine Intelligence, 16(9):920–932, September 1994.Google Scholar
- 16.A. Das K.P. Lim, M.N. Chong. A new MRF model for robust estimate of occlusion and motion vector fields. In International Conference on Image Processing, 1997.Google Scholar
- 17.K.G. Lim and R. Prager. Using Markov random field to integrate stereo modules. Technical report, Cambridge University Endineering Department, 1992. Available from http://svr-www.eng.cam.ac.uk/reports.
- 18.J. K. Marroquine. Probabilistic Solution of Inverse Problem. PhD thesis, Massachusetts Institute of Technology, 1985.Google Scholar
- 19.V. Murino, A. Trucco, and C.S. Regazzoni. A probabilistic approach to the coupled reconstruction and restoration of underwater acoustic images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(1):9–22, January 1998.Google Scholar
- 21.S.T. Barnard. Stereo matching. In A.K. Jain and R. Chellappa, editors, Markov Random Fields Theory and Application, pages 245–271. Academic Press, 1993.Google Scholar
- 22.W. Woo and A. Ortega. Stereo image compression with disparity compensation using the MRF model. In Proceedings of Visual Communications and Image Processing (VCIP’96), 1996.Google Scholar
- 23.R. Zabih and J. Woodfill. Non-parametric local transform for computing visual correspondence. In Proceedings of the European Conference on Computer Vision, pages 151–158, Stockholm, 1994.Google Scholar