Surface Reconstruction from Multiple Views Using Apparent Contours and Surface Texture
We describe a novel approach to reconstructing the complete surface of an object from multiple views, where the camera circumnavigates the object. The approach combines the information available from the apparent contour with the information available from the imaged surface texture.
It is demonstrated that this approach of combining two information sources has significant advantages over using either the contour or texture alone: first, the geometric constraints available are complementary, so that the deficiencies of one source can be overcome by the strengths of the other; second, judicious use of the two sources enables an efficient automatic reconstruction algorithm to be developed.
In particular we make the following contributions: it is shown that the set of epipolar tangencies generates an epipolar net at which the visual hull coincides with the surface; a statistical cost function is defined which incorporates terms for error in image intensity similarity and geometric error in the apparent contour; and, an improved photo-consistency constraint is developed for space carving.
Examples of automatically generated texture-mapped graphical models are given for various objects and camera motions. The objects may contain concavities, and have non-trivial topology.
KeywordsCost Function Surface Reconstruction Geometric Error Multiple View Visual Hull
Unable to display preview. Download preview PDF.
- N. Ayache and O. D. Faugeras. Building a consistent 3D representation of a mobile robot environment by combining multiple stereo views. In Proc. International Joint Conference on Artificial Intelligence, pages 808–810, 1987.Google Scholar
- H. H. Baker and T. O. Binford. Depth from edge and intensity based stereo. In IJCAI81, pages 631–636, 1981.Google Scholar
- P. Beardsley, P. Torr, and A. Zisserman. 3D model acquisition from extended image sequences. In Proc. European Conference on Computer Vision, LNCS 1064/1065, pages 683–695. Springer-Verlag, 1996.Google Scholar
- R. Berthilsson and K. Åström. Reconstruction of 3-D curves from 2-D images using affine shape methods for curves. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, 1997.Google Scholar
- R. Cipolla, K. Åström, and P. Giblin. Motion from the frontier of curved surfaces. In Proc. 5th International Conference on Computer Vision, Boston, pages 269–275, 1995.Google Scholar
- G. Cross and A. Zisserman. Quadric surface reconstruction from dual-space geometry. In Proc. 6th International Conference on Computer Vision, Bombay, India, pages 25–31, January 1998.Google Scholar
- O. Faugeras and R. Keriven. Complete dense stereovision using level set methods. In Proc. 5th European Conference on Computer Vision, Freiburg, Germany, pages 379–393, 1998.Google Scholar
- A. W. Fitzgibbon and A. Zisserman. Automatic camera recovery for closed or open image sequences. In Proc. European Conference on Computer Vision, pages 311–326. Springer-Verlag, June 1998.Google Scholar
- P. Giblin and R. Weiss. Reconstruction of surfaces from profiles. In Proc. 1st International Conference on Computer Vision, London, pages 136–144, London, 1987.Google Scholar
- W. E. L. Grimson. From Images to Surfaces: A Computational Study of the Human Early Visual System. MIT Press, 1981.Google Scholar
- T. Joshi, N. Ahuja, and J. Ponce. Structure and motion estimation from dynamic silhouettes under perspective projection. In Proc. 5th International Conference on Computer Vision, Boston, pages 290–295, 1995.Google Scholar
- R. Koch. 3D surface reconstruction from stereoscopic image sequences. In Proc. 5th International Conference on Computer Vision, Boston, pages 109–114, 1995.Google Scholar
- K. Kutulakos. Affine surface reconstruction by purposive viewpoint control. In Proc. 5th International Conference on Computer Vision, Boston, pages 894–901, 1995.Google Scholar
- K. Kutulakos and S. Seitz. What do N photographs tell us about 3D shape? Technical Report 680, University of Rochester, January 1998.Google Scholar
- T. McInerney and D. Terzopoulos. Topologically adaptable snakes. In Proc. 5th International Conference on Computer Vision, Boston, pages 840–845, 1995.Google Scholar
- S.M. Seitz and C.R. Dyer. Photorealistic scene reconstruction by voxel coloring. In Proc. IEEE Conference on Computer Vision and Pattern Recognition, Puerto Rico, pages 1067–1073, 1997.Google Scholar
- J. A. Sethian. Level Set Methods. Cambridge University Press, Cambridge, 1998.Google Scholar
- A. Shashua and S. Toelg. The quadric reference surface: Theory and applications. In Proc. International Conference on Computer Vision, 1997.Google Scholar