Instantiating deformable models with a neural net

Christopher K. I. Williams; Michael Revow; Geoffrey E. Hinton

doi:10.1006/cviu.1997.0540

Instantiating deformable models with a neural net

Christopher K. I. Williams, Michael Revow, Geoffrey E. Hinton

Research output: Contribution to journal › Article › peer-review

Abstract

Deformable models are an attractive approach to recognizing objects which have considerable within-class variability such as handwritten characters. However, there are severe search problems associated with fitting the models to data which could be reduced if a better starting point for the search were available. We show that by training a neural network to predict how a deformable model should be instantiated from an input image, such improved starting points can be obtained. This method has been implemented for a system that recognizes handwritten digits using deformable models, and the results show that the search time can be significantly reduced without compromising recognition performance. © 1997 Academic Press.

Original language	English
Pages (from-to)	120-126
Number of pages	7
Journal	Computer Vision and Image Understanding
Volume	68
Issue number	1
DOIs	https://doi.org/10.1006/cviu.1997.0540
Publication status	Published - 1 Oct 1997

Bibliographical note

NOTICE: this is the author’s version of a work that was accepted for publication in Computer Vision and Image Understanding. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Williams, Christopher K. I.; Revow, Michael and Hinton, Geoffrey E. (1997). Instantiating deformable models with a neural net. Computer Vision and Image Understanding, 68 (1), 120-126. DOI http://dx.doi.org/10.1006/cviu.1997.0540

Keywords

deformable models
within-class variability
handwritten digits

Access to Document

10.1006/cviu.1997.0540

Instantiating deformable models with a neural net
NOTICE: this is the author’s version of a work that was accepted for publication in Computer Vision and Image Understanding. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Williams, Christopher K. I.; Revow, Michael and Hinton, Geoffrey E. (1997). Instantiating deformable models with a neural net. Computer Vision and Image Understanding, 68 (1), 120-126. DOI http://dx.doi.org/10.1006/cviu.1997.0540
Accepted author manuscript, 198 KB

Cite this

@article{765a5aee0d3b46f4b4e07d236a98a925,

title = "Instantiating deformable models with a neural net",

abstract = "Deformable models are an attractive approach to recognizing objects which have considerable within-class variability such as handwritten characters. However, there are severe search problems associated with fitting the models to data which could be reduced if a better starting point for the search were available. We show that by training a neural network to predict how a deformable model should be instantiated from an input image, such improved starting points can be obtained. This method has been implemented for a system that recognizes handwritten digits using deformable models, and the results show that the search time can be significantly reduced without compromising recognition performance. {\textcopyright} 1997 Academic Press.",

keywords = "deformable models, within-class variability, handwritten digits",

author = "Williams, {Christopher K. I.} and Michael Revow and Hinton, {Geoffrey E.}",

note = "NOTICE: this is the author{\textquoteright}s version of a work that was accepted for publication in Computer Vision and Image Understanding. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Williams, Christopher K. I.; Revow, Michael and Hinton, Geoffrey E. (1997). Instantiating deformable models with a neural net. Computer Vision and Image Understanding, 68 (1), 120-126. DOI http://dx.doi.org/10.1006/cviu.1997.0540",

year = "1997",

month = oct,

day = "1",

doi = "10.1006/cviu.1997.0540",

language = "English",

volume = "68",

pages = "120--126",

journal = "Computer Vision and Image Understanding",

issn = "1077-3142",

publisher = "Academic Press Inc.",

number = "1",

}

TY - JOUR

T1 - Instantiating deformable models with a neural net

AU - Williams, Christopher K. I.

AU - Revow, Michael

AU - Hinton, Geoffrey E.

N1 - NOTICE: this is the author’s version of a work that was accepted for publication in Computer Vision and Image Understanding. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Williams, Christopher K. I.; Revow, Michael and Hinton, Geoffrey E. (1997). Instantiating deformable models with a neural net. Computer Vision and Image Understanding, 68 (1), 120-126. DOI http://dx.doi.org/10.1006/cviu.1997.0540

PY - 1997/10/1

Y1 - 1997/10/1

N2 - Deformable models are an attractive approach to recognizing objects which have considerable within-class variability such as handwritten characters. However, there are severe search problems associated with fitting the models to data which could be reduced if a better starting point for the search were available. We show that by training a neural network to predict how a deformable model should be instantiated from an input image, such improved starting points can be obtained. This method has been implemented for a system that recognizes handwritten digits using deformable models, and the results show that the search time can be significantly reduced without compromising recognition performance. © 1997 Academic Press.

AB - Deformable models are an attractive approach to recognizing objects which have considerable within-class variability such as handwritten characters. However, there are severe search problems associated with fitting the models to data which could be reduced if a better starting point for the search were available. We show that by training a neural network to predict how a deformable model should be instantiated from an input image, such improved starting points can be obtained. This method has been implemented for a system that recognizes handwritten digits using deformable models, and the results show that the search time can be significantly reduced without compromising recognition performance. © 1997 Academic Press.

KW - deformable models

KW - within-class variability

KW - handwritten digits

UR - http://www.scopus.com/inward/record.url?scp=0031257313&partnerID=8YFLogxK

UR - https://www.sciencedirect.com/science/article/pii/S1077314297905403?via%3Dihub

U2 - 10.1006/cviu.1997.0540

DO - 10.1006/cviu.1997.0540

M3 - Article

SN - 1077-3142

VL - 68

SP - 120

EP - 126

JO - Computer Vision and Image Understanding

JF - Computer Vision and Image Understanding

IS - 1

ER -

Instantiating deformable models with a neural net

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this