The geometry and matching of curves in multiple views
In this paper there are two innovations. First, the geometry of imaged curves is developed in two and three views. A set of results are given for both conics and non-algebraic curves. It is shown that the homography between the images induced by the plane of the curve can be computed from two views given only the epipolar geometry, and that the trifocal tensor can be used to transfer a conic or the curvature from two views to a third.
The second innovation is an algorithm for automatically matching individual curves between images. The algorithm uses both photometric information and the multiple view geometric results. For image pairs the homography facilitates the computation of a neighbourhood cross-correlation based matching score for putative curve correspondences. For image triplets cross-correlation matching scores are used in conjunction with curve transfer based on the trifocal geometry to disambiguate matches. Algorithms are developed for both short and wide baselines. The algorithms are robust to deficiencies in the curve segment extraction and partial occlusion.
Experimental results are given for image pairs and triplets, for varying motions between views, and for different scene types. The method is applicable to curve matching in stereo and trinocular rigs, and as a starting point for curve matching through monocular image sequences.
KeywordsEpipolar Line Epipolar Geometry Curve Match Contour Extraction Trifocal Tensor
Unable to display preview. Download preview PDF.
- 1.P. Beardsley, P. Torr, and A. Zisserman. 3D model acquisition from extended image sequences. In Proc. ECCV, pages 683–695, 1996.Google Scholar
- 2.A. Blake and R. Cipolla. Robust estimation of surface curvature from deformation of apparent contours. In Proc. ECCV, pages 465–474, 1990.Google Scholar
- 3.F.L. Bookstein. Fitting conic sections to scattered data. In Computer Graphics and Image Processing, 9:56–71, 1979.Google Scholar
- 4.A.T. Brint and M. Brady. Stereo matching of curves by least deformation. In International Workshop on Intelligent Robots and Systems, pages 163–170, 1989.Google Scholar
- 5.R. C. K. Chung and R. Nevatia. Use of monocular groupings and occlusion analysis in a hierarchical stereo system. In Proc. CVPR, pages 50–56, 1991.Google Scholar
- 6.G. Cross and A. Zisserman. Quadric surface reconstruction from dual-space geometry. In Proc. ICCV, pages 25–31, 1998.Google Scholar
- 8.R.I. Hartley. A linear method for reconstruction from lines and points. Proc. ICCV, pages 882–887, 1995.Google Scholar
- 10.P. Havaldar and G. Medioni. Segmented shape-descriptions from 3-view stereo. Proc. ICCV, pages 102–108, 1995.Google Scholar
- 11.Q. Luong and T. Vieville. Canonic representations for the geometries of multiple projective views. Technical report, University of California, Berkeley, 1993.Google Scholar
- 14.J.L. Mundy and A. Zisserman, editors. Geometric Invariance in Computer Vision. The MIT Press, Cambridge, MA, USA, 1992.Google Scholar
- 16.S.B. Pollard, J.E.W. Mayhew, and J.P. Frisby. PMF: A stereo correspondence algorithm using a disparity gradient constraint. Perception, 14:449–470, 1985.Google Scholar
- 18.L. Robert and O.D. Faugeras. Curve-based stereo: Figural continuity and curvature. In Proc. CVPR, pages 57–62, 1991.Google Scholar
- 19.R. Safaee-Rad, I. Tchoukanov, B. Benhabib, and K.C. Smith. 3D pose estimation from a quadratic curved feature in two perspective views. In Proc. ICPR, pages 341–344, 1992.Google Scholar
- 20.C. Schmid and A. Zisserman. Automatic line matching across views. In Proc. CVPR, pages 666–671, 1997.Google Scholar
- 21.C. Schmid and A. Zisserman. The geometry and matching of lines and curves over multiple views. Research Report, 1998.Google Scholar
- 22.A. Shashua. Trilinearity in visual recognition by alignment. In Proc. ECCV, pages 479–484, 1994.Google Scholar
- 23.M. Spetsakis and J. Aloimonos. Structure from motion using line correspondences. International Journal of Computer Vision, pages 171–183, 1990.Google Scholar
- 25.R. Vaillant. Using occluding contours for 3D object modeling. In Proc. ECCV, pages 454–464, 1990.Google Scholar
- 28.Zisserman A. and Maybank S. A case against epipolar geometry. In Applications of Invariance in Computer Vision LNCS 825. Springer-Verlag, 1994.Google Scholar