Occlusions, discontinuities, and epipolar lines in stereo
Binocular stereo is the process of obtaining depth information from a pair of left and right views of a scene. We present a new approach to compute the disparity map by solving a global optimization problem that models occlusions, discontinuities, and epipolar-line interactions.
In the model, geometric constraints require every disparity discontinuity along the epipolar line in one eye to always correspond to an occluded region in the other eye, while at the same time encouraging smoothness across epipolar lines. Smoothing coefficients are adjusted according to the edge and junction information. For some well-defined set of optimization functions, we can map the optimization problem to a maximum-flow problem on a directed graph in a novel way, which enables us to obtain a global solution in a polynomial time. Experiments confirm the validity of this approach.
KeywordsUniqueness Constraint Left Image Epipolar Line Matching Edge Occlude Region
Unable to display preview. Download preview PDF.
- 2.N. Ayache. Artificial Vision for Mobile Robots. MIT Press. Cambridge, Mass., 1991.Google Scholar
- 3.P. N. Belhumeur and D. Mumford. A bayesian treatment of the stereo correspondence problem using half-occluded regions. In Proc. IEEE Conf. on Computer Vision and Pattern Recognition, 1992.Google Scholar
- 4.A. Blake and A. Zisserman. Visual Reconstruction. MIT Press, Cambridge, Mass., 1987.Google Scholar
- 5.P. Burt and B. Julesz. A disparity gradient limit for binocular fusion. Science, 208:615–617, 1980.Google Scholar
- 6.B. Cernushi-Frias, D. B. Cooper, Y. P. Hung, and P. Belhumeur. Towards a model-based bayesian theory for estimating and recognizing parameterized 3d objects using two or more images taken from different positions. IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI-11:1028–1052, 1989.CrossRefGoogle Scholar
- 7.A. Champolle, D. Geiger, and S. Mallat. Un algorithme multi-échelle de mise encorrespondance stéréo basé sur les champs markoviens. In 13th GRETSI Conference on Signal and Image Processing, Juan-les-Pins, France, Sept. 1991.Google Scholar
- 8.B. V. Cherkassky and A. V. Goldberg On Implementing Push-Relabel Method for the Maximum Flow Problem. In Proc. 4th Int. Programming and Combinatorial Optimization Conf., pages 157–171, 1995.Google Scholar
- 10.O. Faugeras Three-Dimensional Computer Vision. MIT Press. Cambridge, Mass., 1993.Google Scholar
- 12.B. Gillam and E. Borsting. The role of monocular regions in stereoscopic displays. Perception, 17:603–608, 1988.Google Scholar
- 13.W. E. L. Grimson. From Images to Surfaces. MIT Press. Cambridge, Mass., 1981.Google Scholar
- 14.B. Julesz. Foundations of Cyclopean Perception. The University of Chicago Press, Chicago, 1971.Google Scholar
- 15.T. Kanade and M. Okutomi. A stereo matching algorithm with an adaptive window: theory and experiments. In Proc. Image Understanding Workshop DARPA, PA, September 1990.Google Scholar
- 16.J. Malik. On Binocularly viewed occlusion Junctions. In Fourth European Conference on Computer Vision, vol.1, pages 167–174, Cambridge, UK, 1996. Springer-Verlag.Google Scholar
- 17.D. Marr and T. Poggio. Cooperative computation of stereo disparity. Science, 194:283–287, 1976.Google Scholar
- 21.S. B. Pollard, J. E. W. Mayhew, and J. P. Frisby. Disparity gradients and stereo correspondences. Perception, 1987.Google Scholar
- 22.S. Roy and I. Cox. A Maximum-Flow Formulation of the N-camera Stereo Correspondence Problem In Proc. Int. Conf. on Computer Vision, ICCV'98, Bombay, India 1998.Google Scholar