Precise matching by robust estimation of deformation and local coherence

Zhong Dan Lan; Roger Mohr; Long Quan

doi:10.1007/3-540-63930-6_136

Precise matching by robust estimation of deformation and local coherence

Zhong Dan Lan, Roger Mohr, Long Quan

Research output: Chapter in Book/Conference Proceeding/Report › Conference Paper published in a book › peer-review

Abstract

In this paper, we present a linear method that incorporates information from neighboring pixels for sub-pixel matching. Two algorithms are presented. Both rely on a rough initial estimate of the disparity. The first one is optimized for pairs of images requiring negligible window deformation. The second method is slower but more general and more precise. It is applicable for large window deformation and eliminates false initial matches using robust estimation of the local affine window transformation. The first algorithm attains a precision of 0.05 pixels for interest points and 0.06 for random points in the translational case. For general case, if the deformation is small, the second method gives an accuracy of 0.05 pixels; while for large deformation, it gives an accuracy of about 0.06 pixels for points of interest and 0.10 pixels for random points.

Original language	English
Title of host publication	Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings
Editors	Roland Chin, Ting-Chuen Pong
Publisher	Springer Verlag
Pages	313-320
Number of pages	8
ISBN (Print)	3540639306, 9783540639305
DOIs	https://doi.org/10.1007/3-540-63930-6_136
Publication status	Published - 1997
Externally published	Yes
Event	3rd Asian Conference on Computer Vision, ACCV 1998 - Hong Kong, China Duration: 8 Jan 1998 → 10 Jan 1998

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	1351
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	3rd Asian Conference on Computer Vision, ACCV 1998
Country/Territory	China
City	Hong Kong
Period	8/01/98 → 10/01/98

Bibliographical note

Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 1997.

Keywords

Convolution
Correlation
Matching
Precision
Robustness

Access to Document

10.1007/3-540-63930-6_136

Cite this

Lan, Z. D., Mohr, R., & Quan, L. (1997). Precise matching by robust estimation of deformation and local coherence. In R. Chin, & T.-C. Pong (Eds.), Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings (pp. 313-320). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1351). Springer Verlag. https://doi.org/10.1007/3-540-63930-6_136

Lan, Zhong Dan ; Mohr, Roger ; Quan, Long. / Precise matching by robust estimation of deformation and local coherence. Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings. editor / Roland Chin ; Ting-Chuen Pong. Springer Verlag, 1997. pp. 313-320 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d9ed2e81754d46258fbb54f3e4aef639,

title = "Precise matching by robust estimation of deformation and local coherence",

abstract = "In this paper, we present a linear method that incorporates information from neighboring pixels for sub-pixel matching. Two algorithms are presented. Both rely on a rough initial estimate of the disparity. The first one is optimized for pairs of images requiring negligible window deformation. The second method is slower but more general and more precise. It is applicable for large window deformation and eliminates false initial matches using robust estimation of the local affine window transformation. The first algorithm attains a precision of 0.05 pixels for interest points and 0.06 for random points in the translational case. For general case, if the deformation is small, the second method gives an accuracy of 0.05 pixels; while for large deformation, it gives an accuracy of about 0.06 pixels for points of interest and 0.10 pixels for random points.",

keywords = "Convolution, Correlation, Matching, Precision, Robustness",

author = "Lan, \{Zhong Dan\} and Roger Mohr and Long Quan",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 1997.; 3rd Asian Conference on Computer Vision, ACCV 1998 ; Conference date: 08-01-1998 Through 10-01-1998",

year = "1997",

doi = "10.1007/3-540-63930-6\_136",

language = "English",

isbn = "3540639306",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "313--320",

editor = "Roland Chin and Ting-Chuen Pong",

booktitle = "Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings",

}

Lan, ZD, Mohr, R & Quan, L 1997, Precise matching by robust estimation of deformation and local coherence. in R Chin & T-C Pong (eds), Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1351, Springer Verlag, pp. 313-320, 3rd Asian Conference on Computer Vision, ACCV 1998, Hong Kong, China, 8/01/98. https://doi.org/10.1007/3-540-63930-6_136

Precise matching by robust estimation of deformation and local coherence. / Lan, Zhong Dan; Mohr, Roger; Quan, Long.
Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings. ed. / Roland Chin; Ting-Chuen Pong. Springer Verlag, 1997. p. 313-320 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1351).

Research output: Chapter in Book/Conference Proceeding/Report › Conference Paper published in a book › peer-review

TY - GEN

T1 - Precise matching by robust estimation of deformation and local coherence

AU - Lan, Zhong Dan

AU - Mohr, Roger

AU - Quan, Long

N1 - Publisher Copyright: © Springer-Verlag Berlin Heidelberg 1997.

PY - 1997

Y1 - 1997

N2 - In this paper, we present a linear method that incorporates information from neighboring pixels for sub-pixel matching. Two algorithms are presented. Both rely on a rough initial estimate of the disparity. The first one is optimized for pairs of images requiring negligible window deformation. The second method is slower but more general and more precise. It is applicable for large window deformation and eliminates false initial matches using robust estimation of the local affine window transformation. The first algorithm attains a precision of 0.05 pixels for interest points and 0.06 for random points in the translational case. For general case, if the deformation is small, the second method gives an accuracy of 0.05 pixels; while for large deformation, it gives an accuracy of about 0.06 pixels for points of interest and 0.10 pixels for random points.

AB - In this paper, we present a linear method that incorporates information from neighboring pixels for sub-pixel matching. Two algorithms are presented. Both rely on a rough initial estimate of the disparity. The first one is optimized for pairs of images requiring negligible window deformation. The second method is slower but more general and more precise. It is applicable for large window deformation and eliminates false initial matches using robust estimation of the local affine window transformation. The first algorithm attains a precision of 0.05 pixels for interest points and 0.06 for random points in the translational case. For general case, if the deformation is small, the second method gives an accuracy of 0.05 pixels; while for large deformation, it gives an accuracy of about 0.06 pixels for points of interest and 0.10 pixels for random points.

KW - Convolution

KW - Correlation

KW - Matching

KW - Precision

KW - Robustness

UR - https://openalex.org/W1556368803

UR - https://www.scopus.com/pages/publications/84957380235

U2 - 10.1007/3-540-63930-6_136

DO - 10.1007/3-540-63930-6_136

M3 - Conference Paper published in a book

SN - 3540639306

SN - 9783540639305

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 313

EP - 320

BT - Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings

A2 - Chin, Roland

A2 - Pong, Ting-Chuen

PB - Springer Verlag

T2 - 3rd Asian Conference on Computer Vision, ACCV 1998

Y2 - 8 January 1998 through 10 January 1998

ER -

Lan ZD, Mohr R, Quan L. Precise matching by robust estimation of deformation and local coherence. In Chin R, Pong TC, editors, Computer Vision - ACCV 1998 - 3rd Asian Conference on Computer Vision, Proceedings. Springer Verlag. 1997. p. 313-320. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/3-540-63930-6_136

Precise matching by robust estimation of deformation and local coherence

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this