PDE Based Surface Estimation for Structure from Motion
The method of Faugeras and Keriven for solving the multiple view stereo problem by partial differential equations (PDE’s) in a level set implementation has been well received. The main reasons are that it produces good results and deals effectively with objects which are, topologically, more complex than a ball. This makes it a good choice for the final step in the usual structure from motion approach, namely making a full 3D surface reconstruction from estimates of camera orientations and 3D point structure.
Here an approach is proposed whereby the 3D point structure, estimated using structure from motion, is used to initialize the method of Faugeras and Keriven. The proposed approach has the advantage of a) considerably improving the run time of the approach, and b) making it more resistant towards noisy data and data with patches of low variability. Both advantages make the approach much more effective on real data.
KeywordsTrue Surface Smoothness Constraint Face Data Camera Orientation Usual Structure
- 1.J.A. Bærentzen and H. Aanæs. Computing discrete signed distance fields from triangle meshes. Technical Report 21, IMM, DTU, 2002.Google Scholar
- 2.O. Faugeras and R. Keriven. Variatinal principles, surface evaluation, pde’s, level set methods and the stereo problem. Technical Report 3021, INRIA, October 1996.Google Scholar
- 3.O. Faugeras and R. Keriven. Complete dense stereovision using level set methods. Computer Vision-ECCV’98. 5th European Conference on Computer Vision. Proceedings, pages 379–93 vol.1, 1998.Google Scholar
- 5.O. Faugeras, S. Laveau, L. Robert, G. Csurka, and C. Zeller. 3-d reconstruction of urban scenes from sequences of images. Technical Report 2572, INRIA, June 1995.Google Scholar
- 7.H. Jin, A. Yezzi, and S. Soatto. Variational multiframe stereo in the presence of specular reflections. Technical Report TR01-0017, UCLA, 2001.Google Scholar
- 8.K.N. Kutulakos and S.M. Seitz. A theory of shape by space carving. Computer Vision, 1999. The Proceedings of the Seventh IEEE International Conference on, 1:307–314 vol.1, 1999.Google Scholar
- 9.D. Morris and T. Kanade. Image-consistent surface triangulation. In IEEE Conf. Computer Vision and Pattern Recognition’2000, pages 332–338, 2000.Google Scholar
- 12.S. J. Osher and R. P. Fedkiw. Level Set Methods and Dynamic Implicit Surfaces. Springer Verlag, 1st edition, November 2002.Google Scholar
- 13.B. A. Payne and Arthur W. Toga. Distance field manipulation of surface models. Computer Graphics and Applications, 12(1), 1992.Google Scholar
- 14.M. Pollefeys. Self-calibration and metric 3D reconstruction from uncalibrated image sequences. PhD thesis, K.U.Leuven, 1999.Google Scholar
- 15.J.A. Sethian. Level Set Methods and Fast Marching Methods Evolving Interfaces in Computational Geometry, Fluid Mechanics, Computer Vision, and Materials Science. Cambridge University Press, 1999.Google Scholar
- 16.M. Ziegler, L. Falkenhagen, R. Horst, and D. Kalivas. Evolution of stereoscopic and three-dimensional video. Image Communication, 14, 1998.Google Scholar