A Brute-Force Algorithm for Reconstructing a Scene from Two Projections
(2011) IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2011 p.2961-2968- Abstract
- Is the real problem in finding the relative orientation of two viewpoints the correspondence problem? We argue that this is only one difficulty. Even with known correspondences, popular methods like the eight point algorithm and minimal solvers may break down due to planar scenes or small relative motions. In this paper, we derive a simple, brute-force algorithm which is both robust to outliers and has no such algorithmic degeneracies. Several cost functions are explored including maximizing the consensus set and robust norms like truncated least-squares. Our method is based on parameter search in a four-dimensional space using a new epipolar parametrization. In principle, we do an exhaustive search of parameter space, but the computations... (More)
- Is the real problem in finding the relative orientation of two viewpoints the correspondence problem? We argue that this is only one difficulty. Even with known correspondences, popular methods like the eight point algorithm and minimal solvers may break down due to planar scenes or small relative motions. In this paper, we derive a simple, brute-force algorithm which is both robust to outliers and has no such algorithmic degeneracies. Several cost functions are explored including maximizing the consensus set and robust norms like truncated least-squares. Our method is based on parameter search in a four-dimensional space using a new epipolar parametrization. In principle, we do an exhaustive search of parameter space, but the computations are very simple and easily parallelizable, resulting in an efficient method. Further speedups can be obtained by restricting the domain of possible motions to, for example, planar motions or small rotations. Experimental results are given for a variety of scenarios including scenes with a large portion of outliers. Further, we apply our algorithm to 3D motion segmentation where we outperform state-of-the-art on the well-known Hopkins-155 benchmark database. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/2254574
- author
- Enqvist, Olof LU ; Jiang, Fangyuan LU and Kahl, Fredrik LU
- organization
- publishing date
- 2011
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2011
- pages
- 2961 - 2968
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- conference name
- IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2011
- conference location
- Colorado Springs, CO, United States
- conference dates
- 2011-06-20 - 2011-06-25
- external identifiers
-
- wos:000295615803031
- scopus:80052885912
- ISSN
- 1063-6919
- DOI
- 10.1109/CVPR.2011.5995669
- language
- English
- LU publication?
- yes
- id
- 98547584-2a9b-41b8-ab18-9294bc93a22b (old id 2254574)
- date added to LUP
- 2016-04-01 15:03:28
- date last changed
- 2022-01-28 03:54:07
@inproceedings{98547584-2a9b-41b8-ab18-9294bc93a22b, abstract = {{Is the real problem in finding the relative orientation of two viewpoints the correspondence problem? We argue that this is only one difficulty. Even with known correspondences, popular methods like the eight point algorithm and minimal solvers may break down due to planar scenes or small relative motions. In this paper, we derive a simple, brute-force algorithm which is both robust to outliers and has no such algorithmic degeneracies. Several cost functions are explored including maximizing the consensus set and robust norms like truncated least-squares. Our method is based on parameter search in a four-dimensional space using a new epipolar parametrization. In principle, we do an exhaustive search of parameter space, but the computations are very simple and easily parallelizable, resulting in an efficient method. Further speedups can be obtained by restricting the domain of possible motions to, for example, planar motions or small rotations. Experimental results are given for a variety of scenarios including scenes with a large portion of outliers. Further, we apply our algorithm to 3D motion segmentation where we outperform state-of-the-art on the well-known Hopkins-155 benchmark database.}}, author = {{Enqvist, Olof and Jiang, Fangyuan and Kahl, Fredrik}}, booktitle = {{IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2011}}, issn = {{1063-6919}}, language = {{eng}}, pages = {{2961--2968}}, publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}}, title = {{A Brute-Force Algorithm for Reconstructing a Scene from Two Projections}}, url = {{http://dx.doi.org/10.1109/CVPR.2011.5995669}}, doi = {{10.1109/CVPR.2011.5995669}}, year = {{2011}}, }