Epipolar Geometry from Three Correspondences

In this paper, LO-RANSAC 3-LAF – a new algorithm for the correspondence problem – is described. Exploiting processes proposed for computation of affineinvariant local frames, three point-to-point correspondences are found for each region-to-region correspondence. Consequently, it is sufficient to select only triplets of region correspondences in the hypothesis stage of epipolar geometry estimation by RANSAC. We experimentally show that: 1. LO-RANSAC 3-LAF estimates epipolar geometry in time that is orders of magnitude faster than the standard method, 2. that the precision of the LO-RANSAC 3-LAF and the standard method are comparable, and 3. that RANSAC without local optimisation applied to triplets of points from a single region is significantly less precise than the new LO-RANSAC 3-LAF algorithm. In the experiments, a speed-up factor in orders of thousands is achieved on the problem of epipolar geometry estimation. The proposed method is pushing the limit of solvable problems, allowing EG estimation in correspondence problems with the number of inliers below 10%.

[1]  Stepán Obdrzálek,et al.  Local affine frames for wide-baseline stereo , 2002, Object recognition supported by user interaction for service robots.

[2]  Robert C. Bolles,et al.  Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography , 1981, CACM.

[3]  Richard I. Hartley,et al.  In defence of the 8-point algorithm , 1995, Proceedings of IEEE International Conference on Computer Vision.

[4]  Emanuele Trucco,et al.  Geometric Invariance in Computer Vision , 1995 .

[5]  Bernhard P. Wrobel,et al.  Multiple View Geometry in Computer Vision , 2001 .

[6]  Andrew Zisserman,et al.  MLESAC: A New Robust Estimator with Application to Estimating Image Geometry , 2000, Comput. Vis. Image Underst..

[7]  Andrew Zisserman,et al.  Robust Detection of Degenerate Configurations while Estimating the Fundamental Matrix , 1998, Comput. Vis. Image Underst..

[8]  Andrew Zisserman,et al.  Wide baseline stereo matching , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).

[9]  Luc Van Gool,et al.  Wide Baseline Stereo Matching based on Local, Affinely Invariant Regions , 2000, BMVC.

[10]  Jiri Matas,et al.  Robust wide-baseline stereo from maximally stable extremal regions , 2004, Image Vis. Comput..

[11]  Andrew Zisserman,et al.  Viewpoint invariant texture matching and wide baseline stereo , 2001, Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001.

[12]  Jan Flusser,et al.  Convex Layers: A New Tool for Recognition of Projectively Deformed Point Sets , 1999, CAIP.