Iterative Error Bound Minimisation for AAM Alignment

Jason Sarigih; Roland Goecke

doi:10.1109/ICPR.2006.730

Iterative Error Bound Minimisation for AAM Alignment

Jason Sarigih, Roland Goecke

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

24 Citations (Scopus)

40 Downloads (Pure)

Abstract

The active appearance model (AAM) is a powerful generative method used for modelling and segmenting de-formable visual objects. Linear iterative methods have proven to be an efficient alignment method for the AAM when initialisation is close to the optimum. However, current methods are plagued with the requirement to adapt these linear update models to the problem at hand when the class of visual object being modelled exhibits large variations in shape and texture. In this paper, we present a new precomputed parameter update scheme which is designed to reduce the error bound over the model parameters at every iteration. Compared to traditional update methods, our method boasts significant improvements in both convergence frequency and accuracy for complex visual objects whilst maintaining efficiency

Original language	English
Title of host publication	Proceedings of the 18th international conference on pattern recognition
Editors	Yuan Yan Tang, Patrick Wang
Place of Publication	Australia
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Pages	1192-1195
Number of pages	4
ISBN (Print)	0-7695-2521-0
DOIs	https://doi.org/10.1109/ICPR.2006.730
Publication status	Published - 2006
Externally published	Yes
Event	ICPR2006 - , Hong Kong Duration: 20 Aug 2006 → 24 Aug 2006

Conference

Conference	ICPR2006
Country/Territory	Hong Kong
Period	20/08/06 → 24/08/06

Access to Document

10.1109/ICPR.2006.730

Journal ArticleFinal published version, 82.8 KB

Cite this

@inproceedings{2dc3eda7431a42c0acb5362f8ce1b998,

title = "Iterative Error Bound Minimisation for AAM Alignment",

abstract = "The active appearance model (AAM) is a powerful generative method used for modelling and segmenting de-formable visual objects. Linear iterative methods have proven to be an efficient alignment method for the AAM when initialisation is close to the optimum. However, current methods are plagued with the requirement to adapt these linear update models to the problem at hand when the class of visual object being modelled exhibits large variations in shape and texture. In this paper, we present a new precomputed parameter update scheme which is designed to reduce the error bound over the model parameters at every iteration. Compared to traditional update methods, our method boasts significant improvements in both convergence frequency and accuracy for complex visual objects whilst maintaining efficiency",

author = "Jason Sarigih and Roland Goecke",

year = "2006",

doi = "10.1109/ICPR.2006.730",

language = "English",

isbn = "0-7695-2521-0",

pages = "1192--1195",

editor = "Tang, {Yuan Yan} and Patrick Wang",

booktitle = "Proceedings of the 18th international conference on pattern recognition",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

address = "United States",

note = "ICPR2006 ; Conference date: 20-08-2006 Through 24-08-2006",

}

Iterative Error Bound Minimisation for AAM Alignment. / Sarigih, Jason; Goecke, Roland.
Proceedings of the 18th international conference on pattern recognition. ed. / Yuan Yan Tang; Patrick Wang. Australia: IEEE, Institute of Electrical and Electronics Engineers, 2006. p. 1192-1195.

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

TY - GEN

T1 - Iterative Error Bound Minimisation for AAM Alignment

AU - Sarigih, Jason

AU - Goecke, Roland

PY - 2006

Y1 - 2006

N2 - The active appearance model (AAM) is a powerful generative method used for modelling and segmenting de-formable visual objects. Linear iterative methods have proven to be an efficient alignment method for the AAM when initialisation is close to the optimum. However, current methods are plagued with the requirement to adapt these linear update models to the problem at hand when the class of visual object being modelled exhibits large variations in shape and texture. In this paper, we present a new precomputed parameter update scheme which is designed to reduce the error bound over the model parameters at every iteration. Compared to traditional update methods, our method boasts significant improvements in both convergence frequency and accuracy for complex visual objects whilst maintaining efficiency

AB - The active appearance model (AAM) is a powerful generative method used for modelling and segmenting de-formable visual objects. Linear iterative methods have proven to be an efficient alignment method for the AAM when initialisation is close to the optimum. However, current methods are plagued with the requirement to adapt these linear update models to the problem at hand when the class of visual object being modelled exhibits large variations in shape and texture. In this paper, we present a new precomputed parameter update scheme which is designed to reduce the error bound over the model parameters at every iteration. Compared to traditional update methods, our method boasts significant improvements in both convergence frequency and accuracy for complex visual objects whilst maintaining efficiency

U2 - 10.1109/ICPR.2006.730

DO - 10.1109/ICPR.2006.730

M3 - Conference contribution

SN - 0-7695-2521-0

SP - 1192

EP - 1195

BT - Proceedings of the 18th international conference on pattern recognition

A2 - Tang, Yuan Yan

A2 - Wang, Patrick

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Australia

T2 - ICPR2006

Y2 - 20 August 2006 through 24 August 2006

ER -

Iterative Error Bound Minimisation for AAM Alignment

Abstract

Conference

Access to Document

Fingerprint

Cite this