Citation Context

...hical refinement framework [90, 11, 8, 112]. A univalued representation of the disparity map is also not essential. Multi-valued representations, which can represent several depth values along each line of sight, have been extensively studied recently, especially for large multiview data set. Many of these techniques use a voxel-based representation to encode the reconstructed colors and spatial occupancies or opacities [113, 101, 67, 34, 33, 24]. Another way to represent a scene with more complexity is to use multiple layers, each of which can be represented by a plane plus residual parallax [5, 14, 117]. Finally, deformable surfaces of various kinds have also been used to perform 3D shape reconstruction from multiple images [120, 121, 43, 38]. 3.6. Summary of methods Table 1 gives a summary of some representative stereo matching algorithms and their corresponding taxonomy, i.e., the matching cost, aggregation, and optimization techniques used by each. The methods are grouped to contrast different matching costs (top), aggregation methods (middle), and optimization techniques (third section), while the last section lists some papers outside the framework. As can be seen from this table, quite...

Citation Context

... can be computed very efficiently by max flow algorithms. These methods have been successfully used for a wide variety of vision problems including image restoration [7, 8, 12, 14], stereo and motion =-=[4, 7, 8, 13, 16, 20, 21]-=-, voxel occupancy [23], multicamera scene reconstruction [18] and medical imaging [5, 6, 15]. The output ofsthese algorithms is generally a solution with some interesting theoretical quality guarantee...

Citation Context

...ingle #-expansion of f when for all pixels p # P either f # (p) = f(p) or f # (p) = #. This notion of an expansion was proposed by [7], and forms the basis for several very e#ective stereo algorithms =-=[3, 7, 13, 14]-=-. Our algorithm is very straightforward; we simply select (in a fixed order or at random) a disparity #, and we find the unique configuration within a single #-expansion move (our local improvement st...

Citation Context

...course, the improvement will vary with the percentage of higher-level features successfully matched. Another feature-based approach is to first segment the images and then match the segmented regions =-=[11]-=-, [62]. Birchfield and Tomasi [11] segment stereo images into small planar patches for which correspondence is then determined. As with most feature-based methods, this reduces the match sensitivity t...

Citation Context

...r the ground and facade planes. When there is insufficient information to locate a plane, they sweep a hypothesized plane through space to determine the position which best matches the images. It should be noted here that this method does not allow for any details that deviate from the selected plane. Several stereo algorithms explicitly handle slanted surfaces. Burt et al. (1995) advocate pre-warping the images to a reference plane, such as the ground, before performing binocular stereo. In this way, they achieve greater accuracy as well faster computation due to the reduced disparity range. Birchfield and Tomasi (1999) cast the problem of stereo as image segmentation, which also accounts for affine transformations. Processing iterates between segmentation and affine parameter estimation for each segment using graph cuts (Boykov et al. 2001). Ogale and Aloimonos (2004) proposed an algorithm based on dynamic programming to obtain correspondences between segments of scanlines rather than pixels to account for occlusion. Zabulis and Daniilidis (2004) explicitly addressed non-fronto-parallel surfaces by performing correlation on 3D planes instead of the image plane. Thus, correlation kernels can be aligned with ...

Citation Context

...d to a number of problems [6, 7]. In the realm of stereo, the top contenders for the best stereo shape estimation, on the most common comparison data, either use Belief Propagation [11] or Graph Cuts =-=[3, 5]-=-. Both algorithms allow fast, approximate solutions to MRF’s, which are powerful tools for modelling vision problems, but intractable to solve with reasonable speed until recently. These algorithms ma...

Citation Context

...G84, MMP87, Bar89], nonlinear diffusion of support at different disparity hypotheses [SS98], graph cut methods [RC98, IG98, BVZ99], and the use of graph cuts in conjuction with planar surface fitting =-=[BT99]-=-. 1.1.4 Using layers and regions Many techniques have resorted to using layers to handle scenes with possible textureless regions and large amounts of occlusion. One of the first techniques, in the co...

Citation Context

...gies and EM-style optimization for geometric multi-model fitting. Other geometric model fitting works have used separate elements such as random sampling [35, 26] (as in RANSAC) or EM-style iteration =-=[3]-=-, but none have combined them in a single optimization framework. The experiments in [19] show that their energy-based formulation beats many stateof-the-art algorithms in this area. In other settings...