Better correspondence by registration

From National Research Council Canada

DOIResolve DOI: https://doi.org/10.1007/978-3-642-12297-2_42
AuthorSearch for: ; Search for: 1; Search for:
Name affiliation
  1. National Research Council Canada. NRC Industrial Materials Institute
FormatText
TypeArticle
Proceedings titleComputer Vision: ACCV 2009: revised selected papers. 3, Part III
Series titleLecture Notes in Computer Science; Volume 5996
Conference9th Asian Conference on Computer Vision, September 23-27, 2009, Xi’an
ISSN0302-9743
1611-3349
ISBN978-3-642-12296-5
978-3-642-12297-2
Pages436–447
Abstract
Accurate image correspondence is crucial for estimating multiple-view geometry. In this paper, we present a registration-based method for improving accuracy of the image correspondences. We apply the method to fundamental matrix estimation under practical situations where there are both erroneous matches (outliers) and small feature location errors. Our registration-based method can correct feature locational error to less than 0.1 pixel, remedying localization inaccuracy due to feature detectors. Moreover, we carefully examine feature similarity based on their post-alignment appearance, providing a more reasonable prior for subsequent outlier detection. Experiments show that we can improve feature localization accuracy of the MSER feature detector, which recovers the most accurate feature localization as reported in a recent study by Haja and others. As a result of applying our method, we recover the fundamental matrix with better accuracy and more efficiency.
Publication date
PublisherSpringer
LanguageEnglish
Peer reviewedYes
NPARC number23002569
Export citationExport as RIS
Report a correctionReport a correction
Record identifier1eedffeb-e919-4ccb-b374-fc4cb2f5ec2a
Record created2017-11-30
Record modified2017-11-30
Report a problem on this page
Date modified: