Invariant Descriptors for 3-D Object Recognition and Pose

David Alexander Forsyth, Joseph L. Mundy, Andrew Zisserman, Chris Coelho, Aaron Heller, Charles Rothwell

Research output: Contribution to journalArticle

Abstract

Invariant descriptors are shape descriptors that are unaffected by object pose, by perspective projection, and by the intrinsic parameters of the camera. These descriptors can be constructed using the methods of invariant theory, which are briefly surveyed. A range of applications of invariant descriptors in three-dimensional model-based vision is demonstrated. First, a model-based vision system that recognizes curved plane objects, irrespective of their pose, is demonstrated. Curves are not reduced to polyhedral approximations but are handled as objects in their own right. Models are generated directly from image data. Once objects have been recognized, their pose can be computed. Invariant descriptors for three-dimensional objects with plane faces are described. All these ideas are demonstrated on images of real scenes. The stability of a range of invariant descriptors to measurement error is treated in detail.

Original languageEnglish (US)
Pages (from-to)971-991
Number of pages21
JournalIEEE transactions on pattern analysis and machine intelligence
Volume13
Issue number10
DOIs
StatePublished - Oct 1991
Externally publishedYes

Fingerprint

3D Object Recognition
Object recognition
Descriptors
Invariant
Measurement errors
Model-based
Shape Descriptor
Cameras
Three-dimensional
Invariant Theory
Vision System
Measurement Error
Range of data
Camera
Object
Projection
Face
Curve
Approximation

Keywords

  • Computer vision
  • invariants
  • pose computation
  • recognition

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Computational Theory and Mathematics
  • Artificial Intelligence
  • Applied Mathematics

Cite this

Invariant Descriptors for 3-D Object Recognition and Pose. / Forsyth, David Alexander; Mundy, Joseph L.; Zisserman, Andrew; Coelho, Chris; Heller, Aaron; Rothwell, Charles.

In: IEEE transactions on pattern analysis and machine intelligence, Vol. 13, No. 10, 10.1991, p. 971-991.

Research output: Contribution to journalArticle

Forsyth, David Alexander ; Mundy, Joseph L. ; Zisserman, Andrew ; Coelho, Chris ; Heller, Aaron ; Rothwell, Charles. / Invariant Descriptors for 3-D Object Recognition and Pose. In: IEEE transactions on pattern analysis and machine intelligence. 1991 ; Vol. 13, No. 10. pp. 971-991.
@article{69ab167e08bd44abbd291f1ea40add9e,
title = "Invariant Descriptors for 3-D Object Recognition and Pose",
abstract = "Invariant descriptors are shape descriptors that are unaffected by object pose, by perspective projection, and by the intrinsic parameters of the camera. These descriptors can be constructed using the methods of invariant theory, which are briefly surveyed. A range of applications of invariant descriptors in three-dimensional model-based vision is demonstrated. First, a model-based vision system that recognizes curved plane objects, irrespective of their pose, is demonstrated. Curves are not reduced to polyhedral approximations but are handled as objects in their own right. Models are generated directly from image data. Once objects have been recognized, their pose can be computed. Invariant descriptors for three-dimensional objects with plane faces are described. All these ideas are demonstrated on images of real scenes. The stability of a range of invariant descriptors to measurement error is treated in detail.",
keywords = "Computer vision, invariants, pose computation, recognition",
author = "Forsyth, {David Alexander} and Mundy, {Joseph L.} and Andrew Zisserman and Chris Coelho and Aaron Heller and Charles Rothwell",
year = "1991",
month = "10",
doi = "10.1109/34.99233",
language = "English (US)",
volume = "13",
pages = "971--991",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "10",

}

TY - JOUR

T1 - Invariant Descriptors for 3-D Object Recognition and Pose

AU - Forsyth, David Alexander

AU - Mundy, Joseph L.

AU - Zisserman, Andrew

AU - Coelho, Chris

AU - Heller, Aaron

AU - Rothwell, Charles

PY - 1991/10

Y1 - 1991/10

N2 - Invariant descriptors are shape descriptors that are unaffected by object pose, by perspective projection, and by the intrinsic parameters of the camera. These descriptors can be constructed using the methods of invariant theory, which are briefly surveyed. A range of applications of invariant descriptors in three-dimensional model-based vision is demonstrated. First, a model-based vision system that recognizes curved plane objects, irrespective of their pose, is demonstrated. Curves are not reduced to polyhedral approximations but are handled as objects in their own right. Models are generated directly from image data. Once objects have been recognized, their pose can be computed. Invariant descriptors for three-dimensional objects with plane faces are described. All these ideas are demonstrated on images of real scenes. The stability of a range of invariant descriptors to measurement error is treated in detail.

AB - Invariant descriptors are shape descriptors that are unaffected by object pose, by perspective projection, and by the intrinsic parameters of the camera. These descriptors can be constructed using the methods of invariant theory, which are briefly surveyed. A range of applications of invariant descriptors in three-dimensional model-based vision is demonstrated. First, a model-based vision system that recognizes curved plane objects, irrespective of their pose, is demonstrated. Curves are not reduced to polyhedral approximations but are handled as objects in their own right. Models are generated directly from image data. Once objects have been recognized, their pose can be computed. Invariant descriptors for three-dimensional objects with plane faces are described. All these ideas are demonstrated on images of real scenes. The stability of a range of invariant descriptors to measurement error is treated in detail.

KW - Computer vision

KW - invariants

KW - pose computation

KW - recognition

UR - http://www.scopus.com/inward/record.url?scp=0026243593&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0026243593&partnerID=8YFLogxK

U2 - 10.1109/34.99233

DO - 10.1109/34.99233

M3 - Article

AN - SCOPUS:0026243593

VL - 13

SP - 971

EP - 991

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 10

ER -