E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, March/April 1997.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....GEOMETRY FROM PROFILES UNDER. method based on parameterising the surface by radial curves was developed. Bet ter results can be achieved by using an epipolarparameterisation, together with an interpolation using the osculating circle, as introduced in [27] Further refinements were obtained in [14, 146], and a simple technique was developed in [161] based on a finite difference implementation of [27] Despite its simplicity, the method developed in [161] renders results comparable to those in [14] and [27] and was therefore the technique used here. This work makes use of symmetry properties ....

.... using the osculating circle, as introduced in [27] Further refinements were obtained in [14, 146] and a simple technique was developed in [161] based on a finite difference implementation of [27] Despite its simplicity, the method developed in [161] renders results comparable to those in [14] and [27] and was therefore the technique used here. This work makes use of symmetry properties [165, 95, 167, 37] of the surface of revolution swept out by the rotating object to overcome the main difficulties and drawbacks present in other methods that have attempted to estimate motion from ....

[Article contains additional citation context not shown here]

E. Boyer and M. O. Bergen 3D surface reconstruction using occluding con- tours. Int. Journal of Computer Vision, 22(3):219-233, March/April 1997.

....methods as the ones in [35] 14] and [21] cannot deal with situations where profiles are the only available features in the scene. Earlier attempts to solve the problem of reconstruction from image profiles under known motion include [16, 37, 7] and state of the art algorithms can be found in [34, 3, 39]. We use a simple method based on triangulation to reconstruct the model using the estimated motion. Examples using voxel carving [33, 23] are also shown. Details of the 3D reconstruction of the objects are shown in Fig. 14, Fig. 15 and 27 9.6 5 10 15 20 25 5 10 15 20 25 30 35 image index ....

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219-233, March/April 1997.

....methods as the ones in [35] 14] and [21] cannot deal with situations where profiles are the only available features in the scene. Earlier attempts to solve the problem of reconstruction from image profiles under known motion include [16, 37, 7] and state of the art algorithms can be found in [34, 3, 39]. We use a simple method based on triangulation to reconstruct the model using the estimated motion. Examples using voxel carving [33, 23] are also shown. Details of the 3D reconstruction of the objects are shown in Fig. 14, Fig. 15 and 27 5 10 15 20 25 30 35 0 5 10 15 Image index 5 10 15 ....

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, March/April 1997.

....surface without noticeable texture, point and line correspondences may not be easily established. In this case the profile (apparent contour) of the surface is, very often, the only feature available. This calls for the development of a completely different set of techniques, as those found in [32, 7, 4, 30, 22]. This thesis aims at developing simple and practical techniques for the recovery of 1 CHAPTER 1. INTRODUCTION 2 structure and motion from image profiles, making use of the special properties exhibited by the profiles of a smooth curved object performing circular motion. Such techniques will be ....

....requires that the camera motion to be close to linear, otherwise the 3 correspondent viewing rays will not be close to co planar, and thus will lead to larger errors. Besides, the surface must remain on the same side of the tangents in the projection plane (osculating plane) Boyer and Berger [4] dealt with the reconstruction problem directly from a discrete point of view. Based on the differential analysis developed by Cipolla and Blake, they derived a depth formulation from a local approximation of the surface up to order two for discrete motion. Unlike the osculating circle method, ....

[Article contains additional citation context not shown here]

E. Boyer and M. O. Berger. 3d surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, 1997.

....o#ered limited discriminatory power. In contrast, the proposed approach uses nearly the entire silhouette for recognition. This study builds on geometric insights about the occluding contour and silhouettes of smooth surfaces [7, 11] and their use in determining structure from sequences of images [1, 2, 4, 5, 23]. The basic processing steps for each image include detecting the silhouette curve, computing its dual, and then computing an HD curve (high dimensional curve) which is invariant to rigid transformations. When modeling an object from a camera moving over a trajectory of viewpoints, these ....

E. Boyer and M. Berger. 3d surface reconstruction using occluding contours. Int. J. Computer Vision, 22(3):219--233, 1997.

....a finite number of views. Given the set of views V , the approximation vh V is the best conservative geometry that one can achieve. If one doesn t require a conservative estimate, then better approximations are usually achievable by fitting higher order surface approximations to the observed data [5]. As one uses more and more views, the resulting hull improves, and converges to vh1 . Interestingly though, when one is only concerned with predicting silhouettes and one plans to view the scene from some view v t that lies in the triangle defined by three views v 1 ; v 2 ; v 3 , then the ....

....of the surface geometry. They also analyze of the epipolar correspondence between silhouettes which we use in our silhouette interpolation algorithm. Boyer and Berger show how three discrete views can be used to compute an approximate osculating paraboloid for each point on the contour generator [5]. Many more references are contained in those papers. In computer graphics, silhouette information has been used to enhance the expressive renderings of 3D objects [30, 15, 16] We use the silhouette information to create renderings with the appearance of high resolution. There are also a number ....

BOYER, E., AND BERGER, M. 3d surface reconstruction using occluding contours. IJCV 22, 3 (1997), 219--233.

....of points. In this case the the main feature is the apparent contour, which is the contour of the projected image of the surfaces. 2. Theoretical Background This section summarizes results presented in [5] which are relevant to this work. Under known viewer motion the epipolar parametrization [7, 14, 3], shown in fig. 1, can be used to determine the geometry of the surface in the vicinity of the contour gener (t1) t2) G G r(s,t1) r(s,t2) c(t1) c(t2) Figure 1. Epipolar parametrisation. The surface is parametrised locally by the contour generators from successive viewpoints and ....

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, 1997.

....for smooth surfaces without noticeable texture, point and line correspondences may not be easily established. In this case the profile of the surface is, very often, the only feature available. This calls for the development of a completely different set of techniques, as the ones found in [17, 16, 19, 5, 2, 18]. In this paper we address the problem of structure and motion recovery from the profiles of smooth surfaces. In section 2, the differential geometry of surface will first be briefly reviewed. This forms the theoretical framework for various techniques developed for the reconstruction of surfaces ....

....Vaillant and Faugeras [19] developed a similar algorithm which uses the radial plane instead of the epipolar plane. The osculating circle methods require the camera motion to be close to linear and the surface remains on the same side of the tangents in the projection plane. Boyer and Berger [2] derived a depth formulation from a local approximation of the surface up to order two. Given two corresponding points . and on two successive contour generators with vector positions A and B , by taking the scalar product of the difference with the ....

[Article contains additional citation context not shown here]

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, 1997.

No context found.

E. Boyer, M.-O. Berger, 3D surface reconstruction using occluding contours, International Journal of Computer Vision 22 (3) (1997) 219--233. 130

No context found.

E. Boyer and M.-O. Berger. 3D surface reconstruction using occluding contours. International Journal of Computer Vision, 22(3):219--233, 1997. a. b. c. d.

....contours, however, this method supposes that a complete parametrisation of the surface is a priori available which is not, in general, possible. In this context, we propose a new method which consists of two major phases: 1. We rst reconstruct discrete points on extremal contours. Our approach [8] is based on a local approximation of the surface. This allows a linear estimation of depth to be derived, given three consecutive occluding contours. Unlike other local methods, no assumption is made on the camera motion or on the local surface shape. In addition, local methods present advantages ....

....the regularity of the surface through the minimisation of an energy function. The reconstruction process is described in section 2. In section 3, we present the regularisation procedure. Results on real data are shown in section 4. 2. RECONSTRUCTION In this section, we summarise our method, see [8] for more details. In order to reconstruct a point P belonging to an extremal contour C, three occluding contours are required: the corresponding occluding contour and its two neighbouring contours in the sequence. The epipolar correspondence [8] dene two correspondents of the point P on the ....

[Article contains additional citation context not shown here]

E. Boyer and M.-O. Berger. 3D Surface Reconstruction Using Occluding Contours. Rapport interne 95-R013, Centre de Recherche en Informatique de Nancy, Vand#uvre-l#s-Nancy, January 1995.

....is in the tangent plane of S at P 1 or P 2 . In this situation, points p 1 and p 2 are image projections of the same point P (P = P 1 = P 2 ) and rims of S for camera positions C 1 and C 2 intersect at P . Such a point P is called a multiple point of the sequence considered. It has been shown [Boy 95] that the epipolar correspondence is ambiguous at multiple points. Moreover, epipolar correspondence is extremely noise sensitive in the proximity of these points. Thus, the reconstruction process, which is based on the epipolar correspondence, will be highly biased in the vicinity of multiple ....

E. Boyer and M.-O. Berger. 3D Surface Reconstruction Using Occluding Contours. Rapport interne 95-R-013, Centre de Recherche en Informatique de Nancy, Vand#uvre-l#s-Nancy, January 1995.

....to the surface. In addition, it has been shown that local shape recovery from three or more occluding contours is possible given a known camera motion. Several algorithms [Cip 90, Vai 92, Sze 93] allow such a local reconstruction under the assumption of a linear camera motion. In previous works [Boy 95] we proposed an explicit solution for rim point reconstruction which is correct for any camera motion. The approaches mentioned above are concerned with local shape estimation. However, a complete surface description is needed to build an object model. Seales and Faugeras [Sea 95] generate a ....

....are related to non smoothness of contour distribution in the image. A detection algorithm based on this fact is proposed for plane camera rotations. In this paper, we present a global approach for shape estimation which extends previous results on surface reconstruction from occluding contours [Boy 95] and yields robust surface model. First, we study the case of surface regions where local approximations of the reconstruction method are not valid. This the case, for example, at planar, concave parts of the surface or at surface discontinuities. This corresponds to surface areas which are ....

[Article contains additional citation context not shown here]

E. Boyer and M.-O. Berger. 3D Surface Reconstruction Using Occluding Contours. In CAIP'95, Prague (Czech Republic), September 1995. LNCS, volume 970.

No context found.

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, March/April 1997.

No context found.

E. Boyer and M. O. Berger. 3d surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, March 1997.

No context found.

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, 1997.

No context found.

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, March/April 1997.

No context found.

E. Boyer and M. O. Berger. 3D surface reconstruction using occluding contours. Int. Journal of Computer Vision, 22(3):219--233, March/April 1997.

No context found.

E. Boyer, M. O. Berger, 3d surface reconstruction using occluding contours, Int. Journal of Computer Vision 22 (3) (1997) 219--233.

No context found.

E. Boyer and M.-O. Berger. 3D surface reconstruction using occluding contours. International Journal of Computer Vision, 1997. To appear.

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