Concurrent Stereo Matching: An Image Noise-Driven Model
Most published techniques for reconstructing scenes from stereo pairs follow a conventional strategy of searching for a single surface yielding the best correspondence between the images. The search involves specific constraints on surface continuity, smoothness, and visibility (occlusions) embedded in a matching score – typically an ad hoc linear combination of distinctly different criteria of signal similarity. The coefficients or weighing factors are selected empirically because they dramatically effect accuracy of stereo matching. The single surface assumption is also too restrictive – few real scenes have only one surface.
We introduce a paradigm of concurrent stereo that circumvents in part these problems by separating image matching from a choice of the 3D surfaces. Concurrent stereo matching first detects all likely matching 3D volumes instead of single best matches. Then, starting in the foreground, the volumes are explored, selecting mutually consistent optical surfaces that exhibit high point-wise signal similarity. Local, rather than global, surface continuity and visibility constraints are applied.
Unable to display preview. Download preview PDF.
- 1.Burschka, D., Brown, M.Z., Hager, G.D.: Advances in computational stereo. IEEE Trans. PAMI 25, 993–1008 (2003)Google Scholar
- 4.Hsieh, Y.C., Mckeown Jr., D.M., Perlant, F.P.: Performance evaluation of scene registration and stereo matching for cartographic feature extraction. IEEE Trans. PAMI 14, 214–238 (1992)Google Scholar
- 5.Zitnick, C.L., Kanade, T.: A cooperative algorithm for stereo matching and occlusion detection. IEEE Trans. PAMI 22, 675–684 (2000)Google Scholar
- 8.Ohta, Y., Kanade, T.: Stereo by intra- and inter-scanline search using dynamic programming. IEEE Trans. PAMI 7, 139–154 (1985)Google Scholar
- 11.Boykov, Y., Veksler, O., Zabih, R.: Fast approximate energy minimization via graph cut. IEEE Trans. PAMI 23, 1222–1239 (2001)Google Scholar
- 12.Kim, J., Kolmogorov, V., Zabih, R.: Visual correspondence using energy minimization and mutual information. In: Proc. 9th IEEE Int. Conf. Computer Vision (ICCV 2003), Nice, France, October 13-16, vol. 2, pp. 1003–1010. IEEE Computer Society Press, Los Alamitos (2003)Google Scholar
- 13.Tappen, M.F., Freeman, W.T.: Comparison of graph cuts with belief propagation for stereo, using identical MRF parameters. In: Proc. 9th IEEE Int. Conf. Computer Vision (ICCV 2003), Nice, France, October 13-16, vol. 2, pp. 900–906. IEEE Computer Society Press, Los Alamitos (2003)Google Scholar
- 14.Sun, J., Zheng, N.N., Shum, H.Y.: Stereo matching using belief propagation. IEEE Trans. PAMI 25(7), 787–800 (2003)Google Scholar
- 16.Sun, J., Li, Y., Kang, S.B., Shum, H.Y.: Symmetric stereo matching for occlusion handling. In: Proc. 2005 IEEE Computer Soc. Conf. Computer Vision and Pattern Recognition, San Diego, CA, June 20-26, vol. 2, pp. 399–406. IEEE Computer Society Press, Los Alamitos (2005)Google Scholar
- 18.Liu, J., Gimel’farb, G.: Accuracy of stereo reconstruction by minimum cut, symmetric dynamic programming, and correlation. In: Proc. Image & Vision Computing New Zealand (IVCNZ 2004) Conf., Akaroa, New Zealand, November 21-23, pp. 65–70. Landcare Research, Lincoln (2004)Google Scholar
- 21.Birchfield, S., Tomasi, C.: A pixel dissimilarity measure that is insensitive to image sampling. IEEE Trans. PAMI 20, 401–406 (1998)Google Scholar