Advanced

Outlier Removal Using Duality

Olsson, Carl LU ; Eriksson, Anders P LU and Hartley, Richard (2010) IEEE Int. Conf. on Copmuter Vision and Pattern Recognition In 2010 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) p.1450-1457
Abstract
In this paper we consider the problem of outlier removal for large scale multiview reconstruction problems. An efficient and very popular method for this task is RANSAC. However, as RANSAC only works on a subset of the images, mismatches in longer point tracks may go ndetected. To deal with this problem we would like to have, as a post processing step to RANSAC, a method that works on the entire (or a larger) part of the sequence.



In this paper we consider two algorithms for doing this. The first one is related to a method by Sim & Hartley where a quasiconvex problem is solved repeatedly and the error residuals with the largest error is removed. Instead of solving a quasiconvex problem in each step we show that it... (More)
In this paper we consider the problem of outlier removal for large scale multiview reconstruction problems. An efficient and very popular method for this task is RANSAC. However, as RANSAC only works on a subset of the images, mismatches in longer point tracks may go ndetected. To deal with this problem we would like to have, as a post processing step to RANSAC, a method that works on the entire (or a larger) part of the sequence.



In this paper we consider two algorithms for doing this. The first one is related to a method by Sim & Hartley where a quasiconvex problem is solved repeatedly and the error residuals with the largest error is removed. Instead of solving a quasiconvex problem in each step we show that it is enough to solve a single LP or SOCP which yields a significant speedup. Using duality we show that the same theoretical result holds for our method. The second algorithm is a faster version of the first, and it is related to the popular method of $L_1$-optimization. While it is faster and works very well in practice, there is no theoretical guarantee of success. We show that these two methods are related through duality, and evaluate the methods on a number of data sets with promising results. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
in
2010 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
pages
1450 - 1457
publisher
IEEE--Institute of Electrical and Electronics Engineers Inc.
conference name
IEEE Int. Conf. on Copmuter Vision and Pattern Recognition
external identifiers
  • wos:000287417501063
  • scopus:77956001554
ISSN
1063-6919
ISBN
978-1-4244-6984-0
language
English
LU publication?
yes
id
54ce98ce-087f-4b45-ba8f-c22e495c35ce (old id 1687610)
date added to LUP
2011-07-14 10:09:01
date last changed
2018-06-24 03:59:17
@inproceedings{54ce98ce-087f-4b45-ba8f-c22e495c35ce,
  abstract     = {In this paper we consider the problem of outlier removal for large scale multiview reconstruction problems. An efficient and very popular method for this task is RANSAC. However, as RANSAC only works on a subset of the images, mismatches in longer point tracks may go ndetected. To deal with this problem we would like to have, as a post processing step to RANSAC, a method that works on the entire (or a larger) part of the sequence. <br/><br>
<br/><br>
In this paper we consider two algorithms for doing this. The first one is related to a method by Sim &amp; Hartley where a quasiconvex problem is solved repeatedly and the error residuals with the largest error is removed. Instead of solving a quasiconvex problem in each step we show that it is enough to solve a single LP or SOCP which yields a significant speedup. Using duality we show that the same theoretical result holds for our method. The second algorithm is a faster version of the first, and it is related to the popular method of $L_1$-optimization. While it is faster and works very well in practice, there is no theoretical guarantee of success. We show that these two methods are related through duality, and evaluate the methods on a number of data sets with promising results.},
  author       = {Olsson, Carl and Eriksson, Anders P and Hartley, Richard},
  booktitle    = {2010 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  isbn         = {978-1-4244-6984-0},
  issn         = {1063-6919},
  language     = {eng},
  pages        = {1450--1457},
  publisher    = {IEEE--Institute of Electrical and Electronics Engineers Inc.},
  title        = {Outlier Removal Using Duality},
  year         = {2010},
}