H. H. Baker and T. O. Binford. Depth from edge and intensity based stereo. In Proc. 7th Int. Joint Conf. on Artificial Intelligence, pages 631--636, Vancouver, BC, Canada, August 1981. 177  Home/Search   Document Not in Database   Summary   Related Articles   Check

....While the 2D optimization of Equation (3) can be shown to be NP hard for common classes of smoothness functions [123] dynamic programming can find the global minimum for independent scanlines in polynomial time. Dynamic programming was first used for stereo vision in sparse, edge based methods [3, 83]. More recent approaches have focused on the dense (intensity based) scanline optimization problem [10, 9, 46, 31, 18, 13] These approaches work by computing the minimum cost path through the matrix of all pairwise matching costs between two corresponding scanlines. Partial occlusion is handled ....

H. Baker and T. Binford. Depth from edge and intensity based stereo. In IJCAI81, pages 631--636, 1981.

....The baseline of a stereo pair is the distance 16 between the camera locations of the two images. Disparity refers to the difference in image location between corresponding features in the two images, which is projectively related to the depth of the feature in the scene. Years of research (e.g. [3, 10, 17, 22, 26, 32, 35]) have shown that determining stereo correspondences by computer is difficult problem. In general, current methods are successful only when the images are similar in appearance, as in the case of human vision, which is usually obtained by using cameras that are closely spaced relative to the ....

....In the case where the actual structure and the model coincide at P, p o is projected to P and then reprojected to p k , yielding a correspondence with zero disparity. The fact that the epipolar geometry remains linear after the warping step also facilitates the use of the ordering constraint [3, 13] through a dynamic programming technique. 7.4 The matching algorithm Once the warped offset image is formed, stereo matching proceeds in a straightforward manner between the key and warped offset images. The one complication is that the two images are not rectified in the sense of the epipolar ....

[Article contains additional citation context not shown here]

H. H. Baker and T. O. Binford. Depth from edge and intensity based stereo. In Proceedings of the Seventh IJCAI, Vancouver, BC, pages 631--636, 1981.

....that naturally arises when solving for three camera (trinocular) stereo. Our formulation treats the three cameras symmetrically, while imposing a natural occlusion cost and uniqueness constraint. 1 Introduction Determining shape from stereo has often been posed as a global minimization problem [2, 13]. In these formulations one solves for the shape that best predicts the observed image data and is consistent with a set of prior assumptions. These priors often penalize discontinuities and occlusions in the solutions. Once formulated, the minimization problems are then solved with a variety of ....

H. Baker. Depth from edge and intensity based stereo. PhD thesis, University of Illinois at Urbina Chanmpaign, 1981.

....wire frames in Fig. 23. These, in turn, are converted into the scene model in Fig. 37. Finally, the result of modifying the model in Fig. 36 with a new view is shown in Fig. 43. 3. Stereo Analysis Most stereo matching methods involve matching low level image features, such as image intensities [3, 14, 21, 24] or image edge points [3, 13, 24] Points to be matched may also be chosen as interesting points, e.g. those with high variance in all directions [6, 23] Our method involves matching structural features i.e. junctions extracted from the images. There are several reasons for this. First, ....

....are converted into the scene model in Fig. 37. Finally, the result of modifying the model in Fig. 36 with a new view is shown in Fig. 43. 3. Stereo Analysis Most stereo matching methods involve matching low level image features, such as image intensities [3, 14, 21, 24] or image edge points [3, 13, 24]. Points to be matched may also be chosen as interesting points, e.g. those with high variance in all directions [6, 23] Our method involves matching structural features i.e. junctions extracted from the images. There are several reasons for this. First, feature based matching results in ....

Baker, H.H. and Binford, T.O., Depth from edge and intensity based stereo, in: Proceedings Seventh International Joint Conference on Artificial Intelligence, Vancouver, BC (1981) 631-636.

....Model from Images MARTIN HERMAN, TAKEO KANADE, D SH1GERU KUROE Abstract We describe the current state of the 3 D Mosaic project, whose goal is to incrementally acquire a 3 D model of a complex urban scene from images. The notion of incremental acquisition arises from Manuscript received July 12, 1982; revised April 20, 1983. This work was supported by the Defense Advanced Research Projects Agency, ARPA Order 3597, monitored by the U.S. Air Force Avionics Laboratory under Contract F33615 81 K 1539. M. Herman and T. Kanade are with the Department of Computer Science, Carnegie Mellon ....

....from Images MARTIN HERMAN, TAKEO KANADE, D SH1GERU KUROE Abstract We describe the current state of the 3 D Mosaic project, whose goal is to incrementally acquire a 3 D model of a complex urban scene from images. The notion of incremental acquisition arises from Manuscript received July 12, 1982; revised April 20, 1983. This work was supported by the Defense Advanced Research Projects Agency, ARPA Order 3597, monitored by the U.S. Air Force Avionics Laboratory under Contract F33615 81 K 1539. M. Herman and T. Kanade are with the Department of Computer Science, Carnegie Mellon University, ....

[Article contains additional citation context not shown here]

H. H. Baker and T. O. Binford, "Depth from edge and intensity based stereo," in Proc. IJCAI '81, 1981, pp. 631-636.

....The principal idea of this technique is to minimize a cost function in a bidimensional graph. The research of corresponding points is done subject to conjugated epipolar lines. Order constraints and disparity domain 2 enable the reduction of possible paths and allow the intra line consistency [2]. The advantage of this technique is that it allows the subdivision of the problem of matching in to a set of under problems (restriction to couples of epipolar lines) Each underproblem can be solved globally, thus avoiding error propagation problems on the same line. The principal problem of ....

Baker H. H., Baker T. O., Depth from edge and intensity based stereo. Proc. 7 Int. Joint Confer. On Artificial intelligence, Vancouver, Canada, Aug. 1981.

....may go up and down and have disconnected components due to occlusion, texture edges, etc. Therefore, the space of all possi ble matches is enormous, of size O(MV) A common tech nique for making this optimization problem tractable is to introduce an assumption of depth ordering, or monotonicity [ 1] it i2 . i. 4) With this assumption, Eq. 3) may be solved efficiently using dynamic programming [1, 24, 17, 11, 2, 10, 22, 4] The monotonicity assumption should be used with care, however, since it is violated in the presence of occlusions, and can produce artifacts. In practice, we have ....

....of all possi ble matches is enormous, of size O(MV) A common tech nique for making this optimization problem tractable is to introduce an assumption of depth ordering, or monotonicity [ 1] it i2 . i. 4) With this assumption, Eq. 3) may be solved efficiently using dynamic programming [1, 24, 17, 11, 2, 10, 22, 4]. The monotonicity assumption should be used with care, however, since it is violated in the presence of occlusions, and can produce artifacts. In practice, we have observed that violations of the monotonicity assumption result in dropouts, i.e. portions of the scene that are not reconstructed. ....

[Article contains additional citation context not shown here]

H.H. Baker and T. O. Binford. Depth from edges and intensity based stereo. In lnt. Joint Conf. on Artificial Intelligence, pages 631 636, 1981.

....involved the convolution of a discretized Laplacian of a Gaussian function over the entire image and the edge strength was proportional to the gradient of the convolved output. Other edge detector operation widely used are those proposed by Canny and Deriche [46,47] Others like Baker and Binford [19] and Ohta and Kanade [9] locate peaks of the magnitude of the first derivative intensity profile along an image row and label them as feature points of interest. Gradient edge detectors such as the Roberts, Sobel and Prewitt operators have also had extensive use [25, 35] These operators are of ....

H. H. Baker, and T. O. Binford, "Depth from Edge and Intensity-Based Stereo," Proceedings of the 7th International Joint Conference on Artificial Intelligence, pp. 631-636, Los Altos, California, 1981.

....exceptional circumstances where one can use one of known methods that find a global optimum. For instance, in a situation where the state of the problem is described as a string of linearly ordered local states, dynamic programming can be used (e.g. Amini, Weymouth, and Jain[1] Baker and Binford[2]; Geiger, Gupta, Costa, and Vlontzos[8] Montanari[20] This paper points out another instance of such circumstances and describes a method that can be used; namely, a method to solve a first order Markov Random Field, or MRF, with a prior term that is convex in terms of a linearly ordered label ....

H. H. Baker and T. O. Binford. Depth from Edge and Intensity-based Stereo. In Proceedings of 7th International Joint Conferences on Artificial Intelligence, 631--636, 1981.

....search intra scanlines search [8] This intra scanlines search can be treated as the problem of finding a matching path on a two dimensional (2D) search plane whose vertical and horizontal axes are the right and the left scanlines. A dynamic programming technique can handle this search efficiently [5]. Several factors make the correspondence problem difficult. Indeed, the geometric properties of the picture vary, one from the other because of geometric distortions; moreover objects visibility is often different in the two images [6] 14] The presence of noise or a variation of lighting ....

....But this technique does not consider noise effect. There are some situations in which the images might be corrupted with noise. If the images are severely corrupted by additive correlated Gaussian noise of unknown covariance, Second Order Statistics (SOS) based methods (like dynamic programming [5][13] or block matching. do not work well. In this case, Higher Order Statistics (HOS) based methods are more advantageous since they are not affected by such noise [1] 2] Motivated by the noise insensitivity of Cumulant based estimators, we have proposed a novel correlation method based on a ....

[Article contains additional citation context not shown here]

Baker H. H., Binford T. O., Depth from edge and intensity based stereo. Proc. 7 th Int. Joint Confer. On Artificial intelligence, Vancouver, Canada, Aug. 1981.

....are very numerous. Many methods use some variant of an optimization approach in which the energy functional encourages similar features in the images to correspond, while using a regularizationterm to smooth out noise and to determine the optical flow in regions where the data is non committal [1, 4, 5, 6, 7, 11, 13, 25, 16, 17, 20, 28]. As already discussed, such an approach is complementary to that pursued here. The results of dense computations could be used as extra data for the type of model considered here, as is done in [23] but that is what we are trying to avoid. In fact the reverse is also true: the results of the ....

H. H. Baker and T. O. Binford. Depth from edge and intensity-based stereo. In Proc. 7 th Int'l J'nt Conf. Artif. Intell., pages 631--636 , August 1981.

....correspondence is an important topic in computer vision. It is difficult to obtain correct correspondence by using only local similarity because the search space is huge. Research methods that use general situation similarity are, for example, coarse to fine analysis [6, 4] dynamic programming [2, 8], and structural analysis [7, 1, 5, 3] Thecorrespondence unit of the structural analysis method is a higher order feature, while the units of other methods are basically points. The search space can be narrowed considerably in this method though processing becomes complicated. It is also not ....

H. H. Baker and T. O. Binford. Depth from edge and intensity based stereo. In Proc. of IJICAI81, pp.631-636, 1981.

....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 ....

H.H. Baker and T.O. Binford : "Depth from edges and intensity based stereo," Proc. 7th Int. Joint Conf. Artif. Intell., pp. 631-636, Vancouver, Canada, Aug. 1981.

No context found.

H. H. Baker and T. O. Binford. Depth from edge and intensity based stereo. In Proc. 7th Int. Joint Conf. on Artificial Intelligence, pages 631--636, Vancouver, BC, Canada, August 1981. 177

No context found.

H. H. Baker and T. O. Binford. Depth from Edge and Intensity-based Stereo. In Proceedings of 7th International Joint Conferences on Artificial Intelligence, 631-- 636, 1981.

No context found.

H.H. Baker and T.O. Binford. Depth from edge and intensity based stereo. IJCAI, 631#636, 1981.

No context found.

H.H. Baker, "Depth from Edge and Intensity Based Stereo," Technical Report AIM-347, Artificial Intelligence Laboratory, Stanford Univ., 1982.

No context found.

H. Harlyn Baker and Thomas O. Binford. Depth from edge and intensity-based stereo. In Proceedings of the International Joint Conferences on Artificial Intelligence, pages 631--636, 1981.

No context found.

Baker HH, Binford TO (1981) Depth from Edge and Intensity Based Stereo. In: Proc. 7th Int. Joint Conf. Artificial Intelligence, pp 631--636

No context found.

H. Harlyn Baker and Thomas O. Binford. Depth from edge and intensity-based stereo. In Proc. IJCAI, pages 631--636, 1981.

No context found.

H. H. Baker, T. O. Binford, Depth from edge and intensity based stereo, in: Proc. 7th Int. Joint Conf. on Artificial Intelligence, Vancouver, BC, Canada, 1981, pp. 631--636.

No context found.

Baker H. H., Binford T. O., Depth from edge and intensity based stereo. Proc. 7 Int. Joint Confer. On Artificial intelligence, Vancouver, Canada, Aug. 1981.

No context found.

H. H. Baker and T. O. Binford. Depth from Edge and Intensity-based Stereo. In Proceedings of 7th International Joint Conferences on Artificial Intelligence, 631--636, 1981.

No context found.

Baker, H.H. and Binford, T.O., Depth from edge and intensity based stereo, in: Proceedings Seventh International Joint Conference on Artificial Intelligence, Vancouver, BC (1981

No context found.

H. H. Baker and T. O. Binford, "Depth from Edge and Intensity Based Stereo" in Proc. of 7th IJCAI, vol. 2, pp. 631 636, 1981.

First 50 documents  Next 50

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