The correspondence problem in stereo vision is notoriously difficult. In many approaches a noisy solution is extracted from the correspondence space. Various sophisticated regularization techniques are applied then on this noisy solution. We study here the possibility to denoise the correspondence/correlation space before extracting the solution, by a non-linear and non-isotropic scheme. We show that this methods preserves edges (depth discontinuities) well and overcomes some of the problems encountered in previous approaches.
|Number of pages||4|
|Journal||Proceedings - International Conference on Pattern Recognition|
|State||Published - 2004|
|Event||Proceedings of the 17th International Conference on Pattern Recognition, ICPR 2004 - Cambridge, United Kingdom|
Duration: 23 Aug 2004 → 26 Aug 2004