| W. Hoff and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. IEEE Trans. on Pattern Analysis and Machine Intelligence, pages 121--136, 1989. |
....explicitly, they may incorrectly match adjacent parts and interfere with the correct matching in the neighborhood of occluded regions. Thus a matching process taking into account occlusions is necessary to accurately recover the 3 D structure from 2 D stereo images. Recently several authors [27] [14] [6] 5] 19] proposed some new computational frameworks for stereo matching incorporating occlusion information. A common characteristic among these methods is that two matching processes (L to R and R to L) run separately and benefit little from each other. A way for further improvement of ....
W.Hoff and N.Ahuja, Surface from stereo: Integrating feature matching, disparity estimation and contour detection, IEEE Trans. Pattern Anal. Machine Intell., Vol. 11, 1989
....search. In this paper we present a method that improves the performance of stereo matching in the metric recovery of facial shape by using active shape model detected facial features. Two broad classes of techniques are used in stereo vision; area based [5 9] and feature based matching [10 14]. A feature based approach provides a sparse disparity map by matching at texture rich points only. By itself, it does not provide sufficient information for facial recovery. Area based stereo matching chooses an intensity BMVC99 533 window in one image and then searches for its homologous ....
Hoff W 1989 Surface from Stereo: Integrating Feature Matching, Disparity Estimation, and Contour Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence. 11(2), 121-136.
....is chosen. This method ignores the issue of window shape. Thus it cannot work well near disparity discontinuities in regions of low texture, since in such cases a large window not centered in the middle is needed. The importance of edges for the stereo correspondence has been frequently argued [10, 4], since intensity edges frequently coincide with depth discontinuity edges. Thus some authors 6 chose rely on the intensity edge information to construct windows which do not overlap disparity discontinuity. For example Lotti and Giraudon [18] use the smallest edge limited contour window between ....
W. Ho and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. PAMI, 11(2):121-136, February 1989.
....are one of the major sources for wrong matches. Most of the recent stereo and optical flow work consists of incremental improvements to existing methods, to increase speed, accuracy or reliability. Only a few authors directly treat large occlusion stereo [IB94] Usually, coarse to fine (e.g. [HA89]) or hierarchical (e.g. MT94] matching strategies seems to be necessary to deal with large disparity range. Our algorithm uses mainly region growing techniques. Region growing is a classic approach for segmentation [HS85] Mon87] and finding shapes [Bra93] In its simplest sense, region ....
W. Hoff and N. Ahuja. Surfaces from Stereo: Integrating Feature Matching, Disparity Estimation, and Contour Detection. PAMI, 11(2): 121--136, 1989.
.... based methods use sets of pixels with similar attributes, normally either pixels belonging to edges Z Kim and Aggarwal, 1987; Marr and Poggio, 1979; Mousavi and Schalkoff, 1994; Pollard et al. 1981 Z or the corresponding edges themselves Ayache and Faverjon, 1987; Cruz et al. 1995a,b; Hoff and Ahuja, 1989; Medioni and Nevatia, 1985; Pajares, 1995 . As Z. shown in Ozanian, 1995 , these last methods lead to a sparse depth map only, leaving the rest of the surface to be reconstructed by interpolation; but they are faster than area based methods, because there are Z. much fewer points features to be ....
....high reliability Breuel, 1996 Z.Z. and robustness Wuescher and Boyer, 1991 , b they are abundant in the environment where the experi Z. ments have been carried out, and c they have been successfully used in previous stereovision matching Z works Ayache and Faverjon, 1987; Cruz et al. 1995a,b; Hoff and Ahuja, 1989; Medioni and Neva . tia, 1985; Pajares, 1995 . The contour edges in both images are extracted using the Laplacian of Gaussian filter in accordance Z with the zero crossing criterion Huertas and . Medioni, 1986 . For each zero crossing in a given Z. image, its gradient vector magnitude and ....
Hoff, W., Ahuja, N., 1989. Surface from stereo: Integrating feature matching, disparity estimation and contour detection. IEEE Trans. Pattern Anal. Machine Intell. 11, 121--136.
....scene radiance all contribute to a photograph s dependence on viewpoint. Since our notion of photo consistency implicitly ensures that all of these 3D shape cues are taken into account in the recovery process, our approach is related to work on stereo (Okutomi and Kanade, 1993; Cox et al. 1996; Hoff and Ahuja, 1989), shape fromcontour (Cipolla and Blake, 1992; Vaillant and Faugeras, 1992; Szeliski, 1993) as well as shape from shading (Epstein et al. 1996; Belhumeur and Kriegman, 1996; Woodham et al. 1991) These approaches rely on studying a single 3D shape cue under the assumptions that other sources of ....
Hoff, W. and N. Ahuja: 1989, `Surfaces from Stereo: Integrating Feature Matching, Disparity Estimation, and Contour Detection'. IEEE Trans. Pattern Anal. Machine Intell. 11, 121--136.
....areas small windows are inherently ambiguous. A common strategy for combating this problem is to enforce depth smoothness using techniques such as cooperative competitive algorithms [Marr79, Zitnick99] multiresolution schemes [Hanna93] graph methods [Roy98, Boykov99] and surface model fitting [Hoff89] However, some of these techniques may introduce excessive smoothness that blurs the depth discontinuities and fails to capture details of the scene such as thin structures. Most of these algorithms face the classic depth smoothness and accuracy tradeoff in one way or the other. Using adaptive ....
W. Hoff and N. Ahuja, "Surfaces from stereo: integrating feature matching, disparity estimation, and contour detection," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 11, no. 2, pp. 121-136, February 1989.
....first step by feeding back additional constraints, such as smoothness, that can be useful in producing better results from the first step. In order to address the above problem, there have been a number of proposals to integrate the main analysis phase with the extraction of 3D data. Ho# and Ahuja[11] and Fua[9] combine the steps of stereo matching and surface reconstruction. Kambhamettu et al. 13] couple motion estimation analysis with stereo matching problem. Faugeras and Keriven[8] pose the stereo problem as a variational problem to drive partial di#erential equations, which are solved by ....
....Following the research trend on integration of stereo correspondence and surface reconstruction, in this paper we present a novel method to unify stereo correspondence, continuous surface reconstruction and tracking at the same step using a deformable dual mesh. Although we find Ho# and Ahuja[11], Fua[9] and Faugeras and Keriven[8] closest to our work, our system is fundamentally very di#erent in the assumptions and in the basic methods used. The basic similarity between the above three systems and our system is that all four of them are formulated as an optimization framework. Ho# and ....
[Article contains additional citation context not shown here]
W. Ho# and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. PAMI, 11(2):121--136, February 1989.
....used is a 85MHz Sun SPARCserver1000 running Solaris 2.5. The 14 (a) b) c) d) e) f) Fig. 11: The matching result for Random Dot Stereograms. a,d) left image; b,e) right image; and (c,f) the disparity map recovered. Images (d,e) courtesy of Bill Hoff at the University of Illinois [30]. a) b) c) d) e) f) Fig. 12: The matching result for synthetic images. The image sizes are 256 Theta256. Top row shows a two level background. Bottom row are the images of a sphere on a table. a,d) left image; b,e) right image; and (c,f) the disparity map recovered. Images ....
....for synthetic images. The image sizes are 256 Theta256. Top row shows a two level background. Bottom row are the images of a sphere on a table. a,d) left image; b,e) right image; and (c,f) the disparity map recovered. Images (a,b,d,e) courtesy of Bill Hoff at the University of Illinois [30]. 15 (a) b) c) d) e) Fig. 13: The matching result for a real image ( ball ) a) left image; b) right image; c) the disparity map recovered. d) the perspective view of the result with the original image draped over the disparity map; and (e) occlusion detected. Images (a,b) ....
[Article contains additional citation context not shown here]
W. Hoff and N. Ahuja, "Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 11, pp. 121--136, February 1989. 24
....fig.3) This circumstance is taken into account during estimation by defining an additional disparity offset d off , and must also be treated during interpolation (see section V.2) During the last years, many different schemes for disparity estimation have been proposed. Though feature based [4,5,6] and dynamic programming [7,8,9] approaches seem to perform very well, we found them to be too complex for a hardware system with the requirement of large disparity ranges even in the case of pure horizontal disparities. Matching approaches can be classified as area based schemes [10,11] We have ....
W. Hoff and N. Ahuja : "Surfaces from stereo : Integrating feature matching, disparity estimation and contour detection," IEEE Trans. Patt Anal. Mach. Intell., vol. PAMI-11, no.2, 1989.
....locally smooth surface reconstruction in the sense that the residual between detected and modeled disparity is as small as possible. Yet another approach for surface modeling with special emphasis on the integration of feature matching, contour detection, and surface interpolation is presented in [10]. At the first stage, a planar model is used as a local approximation of disparity. Then a quadratic model is fitted to larger patches which are formed by comparing the mutual consistency of the neighboring planar patches. Ridge and occluding contours are detected by fitting bipartite circular ....
W. Hoff and N. Ahuja, "Surfaces from stereo: integrating feature matching, disparity estimation, and contour detection", IEEE Trans. Pattern Anal. Machine Intell., Vol. 11, pp. 121-136, February 1989.
No context found.
W. Hoff and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. IEEE Trans. on Pattern Analysis and Machine Intelligence, pages 121--136, 1989.
No context found.
W. Ho# and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 11(2):121--136, February 1989.
No context found.
W. Hoff and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 11(2):121--136, February 1989.
No context found.
W. Hoff and N. Ahuja, "Surfaces from Stereo: Integrating Feature Matching, Disparity Estimation, and Contour Detection," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 11, no. 2, pp. 121-136, 1989.
No context found.
W. Hoff and N. Ahuja. Surfaces from stereo: integrating feature matching, disparity estimation, and contour detection. IEEE Trans. Pat. Anal. Mach. Int., 11(2):121--136, 1989.
No context found.
W. Ho# and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 11(2):121--136, 1989.
No context found.
W. Hoff and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. IEEE Trans. PAMI, 11(2):121--136, 1989.
No context found.
W. Hoff and N. Ahuja, "Surfaces from stereo: Integrating feature matching, disparity estimation and contour detection," IEEE Trans. Pattern Anal. Machine Intell., vol. 11, pp. 121--136, Feb. 1989.
No context found.
W. Ho# and N. Ahuja. Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 11(2):121--136, 1989.
No context found.
W. Hoff and N. Ahuja, "Surfaces from stereo: Integrating feature matching, disparity estimation, and contour detection, " IEEE Trans. Pattern Anal. Machine Intell., vol. 11, pp. 121--136, 1989. 17
No context found.
W. Hoff and N. Ahuja, "Surfaces from Stereo: Integrating Feature Matching, Disparity Estimation, and Contour Detection", IEEE Trans. on PAMI, vol. 11, no. 2, pp. 121-136, 1989.
No context found.
# W. Hoff and N. Ahuja, "Surfaces from Stereo: Integrating Feature Matching, Disparity Estimation, and Contour Detection," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 11, no. 2, pp. 121136, Feb. 1989.
No context found.
Ho#, W. and Ahuja, N., #Surfaces from Stereo: Integrating Feature Matching, Disparity Estimation, and Contour Detection," IEEE Trans. Patt. Anal. Mach. Intell. 11, pp. 121#136, February 1989.
No context found.
W. Hoff and N. Ahuja : "Surfaces from stereo : Integrating feature matching, disparity estimation and contour detection," IEEE Trans. Patt Anal. Mach. Intell., vol. PAMI-11, no.2, 1989
First 50 documents
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC