Recognising rotationally symmetric surfaces from their outlines

David A. Forsyth, Joseph L. Mundy, Andrew Zisserman, Charles A. Rothwell

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Recognising a curved surface from its outline in a single view is a major open problem in computer vision. This paper shows techniques for recognising a significant class of surfaces from a single perspective view. The approach uses geometrical facts about bitangencies, creases, and inflections to compute descriptions of the surface’s shape from its image outline. These descriptions are unaffected by the viewpoint or the camera parameters. We show, using images of real scenes, that these representations identify surfaces from their outline alone. This leads to fast and effective recognition of curved surfaces. The techniques we describe work for surfaces that have a rotational symmetry, or are projectively equivalent to a surface with a rotational symmetry, and can be extended to an even larger class of surfaces. All the results in this paper are for the case of full perspective. The results additionally yield techniques for identifying the line in the image plane corresponding to the axis of a rotationally symmetric surface, and for telling whether a surface is rotationally symmetric or not from its outline alone.

Original languageEnglish (US)
Title of host publicationComputer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings
EditorsGiulio Sandini
PublisherSpringer-Verlag
Pages639-647
Number of pages9
ISBN (Print)9783540554264
DOIs
StatePublished - Jan 1 1992
Event2nd European Conference on Computer Vision, ECCV 1992 - Santa Margherita Ligure, Italy
Duration: May 19 1992May 22 1992

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume588 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other2nd European Conference on Computer Vision, ECCV 1992
CountryItaly
CitySanta Margherita Ligure
Period5/19/925/22/92

Fingerprint

Rotational symmetry
Curved Surface
Computer Vision
Open Problems
Camera
Computer vision
Cameras
Line
Class

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Forsyth, D. A., Mundy, J. L., Zisserman, A., & Rothwell, C. A. (1992). Recognising rotationally symmetric surfaces from their outlines. In G. Sandini (Ed.), Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings (pp. 639-647). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 588 LNCS). Springer-Verlag. https://doi.org/10.1007/3-540-55426-2_68

Recognising rotationally symmetric surfaces from their outlines. / Forsyth, David A.; Mundy, Joseph L.; Zisserman, Andrew; Rothwell, Charles A.

Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings. ed. / Giulio Sandini. Springer-Verlag, 1992. p. 639-647 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 588 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Forsyth, DA, Mundy, JL, Zisserman, A & Rothwell, CA 1992, Recognising rotationally symmetric surfaces from their outlines. in G Sandini (ed.), Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 588 LNCS, Springer-Verlag, pp. 639-647, 2nd European Conference on Computer Vision, ECCV 1992, Santa Margherita Ligure, Italy, 5/19/92. https://doi.org/10.1007/3-540-55426-2_68
Forsyth DA, Mundy JL, Zisserman A, Rothwell CA. Recognising rotationally symmetric surfaces from their outlines. In Sandini G, editor, Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings. Springer-Verlag. 1992. p. 639-647. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/3-540-55426-2_68
Forsyth, David A. ; Mundy, Joseph L. ; Zisserman, Andrew ; Rothwell, Charles A. / Recognising rotationally symmetric surfaces from their outlines. Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings. editor / Giulio Sandini. Springer-Verlag, 1992. pp. 639-647 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{c3ec6c6a7e334f249ed05b03477a19b3,
title = "Recognising rotationally symmetric surfaces from their outlines",
abstract = "Recognising a curved surface from its outline in a single view is a major open problem in computer vision. This paper shows techniques for recognising a significant class of surfaces from a single perspective view. The approach uses geometrical facts about bitangencies, creases, and inflections to compute descriptions of the surface’s shape from its image outline. These descriptions are unaffected by the viewpoint or the camera parameters. We show, using images of real scenes, that these representations identify surfaces from their outline alone. This leads to fast and effective recognition of curved surfaces. The techniques we describe work for surfaces that have a rotational symmetry, or are projectively equivalent to a surface with a rotational symmetry, and can be extended to an even larger class of surfaces. All the results in this paper are for the case of full perspective. The results additionally yield techniques for identifying the line in the image plane corresponding to the axis of a rotationally symmetric surface, and for telling whether a surface is rotationally symmetric or not from its outline alone.",
author = "Forsyth, {David A.} and Mundy, {Joseph L.} and Andrew Zisserman and Rothwell, {Charles A.}",
year = "1992",
month = "1",
day = "1",
doi = "10.1007/3-540-55426-2_68",
language = "English (US)",
isbn = "9783540554264",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer-Verlag",
pages = "639--647",
editor = "Giulio Sandini",
booktitle = "Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings",

}

TY - GEN

T1 - Recognising rotationally symmetric surfaces from their outlines

AU - Forsyth, David A.

AU - Mundy, Joseph L.

AU - Zisserman, Andrew

AU - Rothwell, Charles A.

PY - 1992/1/1

Y1 - 1992/1/1

N2 - Recognising a curved surface from its outline in a single view is a major open problem in computer vision. This paper shows techniques for recognising a significant class of surfaces from a single perspective view. The approach uses geometrical facts about bitangencies, creases, and inflections to compute descriptions of the surface’s shape from its image outline. These descriptions are unaffected by the viewpoint or the camera parameters. We show, using images of real scenes, that these representations identify surfaces from their outline alone. This leads to fast and effective recognition of curved surfaces. The techniques we describe work for surfaces that have a rotational symmetry, or are projectively equivalent to a surface with a rotational symmetry, and can be extended to an even larger class of surfaces. All the results in this paper are for the case of full perspective. The results additionally yield techniques for identifying the line in the image plane corresponding to the axis of a rotationally symmetric surface, and for telling whether a surface is rotationally symmetric or not from its outline alone.

AB - Recognising a curved surface from its outline in a single view is a major open problem in computer vision. This paper shows techniques for recognising a significant class of surfaces from a single perspective view. The approach uses geometrical facts about bitangencies, creases, and inflections to compute descriptions of the surface’s shape from its image outline. These descriptions are unaffected by the viewpoint or the camera parameters. We show, using images of real scenes, that these representations identify surfaces from their outline alone. This leads to fast and effective recognition of curved surfaces. The techniques we describe work for surfaces that have a rotational symmetry, or are projectively equivalent to a surface with a rotational symmetry, and can be extended to an even larger class of surfaces. All the results in this paper are for the case of full perspective. The results additionally yield techniques for identifying the line in the image plane corresponding to the axis of a rotationally symmetric surface, and for telling whether a surface is rotationally symmetric or not from its outline alone.

UR - http://www.scopus.com/inward/record.url?scp=85011316617&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85011316617&partnerID=8YFLogxK

U2 - 10.1007/3-540-55426-2_68

DO - 10.1007/3-540-55426-2_68

M3 - Conference contribution

AN - SCOPUS:85011316617

SN - 9783540554264

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

SP - 639

EP - 647

BT - Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings

A2 - Sandini, Giulio

PB - Springer-Verlag

ER -