A model of facial behaviour

F. Bettinger; T. F. Cootes

doi:10.1109/AFGR.2004.1301519

A model of facial behaviour

F. Bettinger
, T. F. Cootes

Research output: Chapter in Book/Conference proceeding › Conference contribution

Abstract

We wish to model the way in which faces move in video sequences. We represent facial behaviour as a sequence of short actions. Each action is a sample from a statistical model representing the variability in the way it is performed. The ordering of actions is defined using a variable length Markov model. Action models and variable length Markov model are trained from a long (20000 frames) video sequence of a talking face. We propose a novel method of quantitatively evaluating the quality of the synthesis by measuring overlaps of parameter histograms. We apply this method to compare our technique with an alternative model that uses an autoregressive process.

Original language	English
Title of host publication	Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition\|Proc. Sixth IEEE Int. Conf. Autom. Face Gest. Recogn.
Publisher	IEEE Computer Society
Pages	123-128
Number of pages	5
ISBN (Print)	0769521223, 9780769521220
DOIs	https://doi.org/10.1109/AFGR.2004.1301519
Publication status	Published - 2004
Event	Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004 - Seoul Duration: 1 Jul 2004 → …

Conference

Conference	Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004
City	Seoul
Period	1/07/04 → …

Access to Document

10.1109/AFGR.2004.1301519

Cite this

@inproceedings{df4de4407c2f4e208c333504c2ecad24,

title = "A model of facial behaviour",

abstract = "We wish to model the way in which faces move in video sequences. We represent facial behaviour as a sequence of short actions. Each action is a sample from a statistical model representing the variability in the way it is performed. The ordering of actions is defined using a variable length Markov model. Action models and variable length Markov model are trained from a long (20000 frames) video sequence of a talking face. We propose a novel method of quantitatively evaluating the quality of the synthesis by measuring overlaps of parameter histograms. We apply this method to compare our technique with an alternative model that uses an autoregressive process.",

author = "F. Bettinger and Cootes, \{T. F.\}",

year = "2004",

doi = "10.1109/AFGR.2004.1301519",

language = "English",

isbn = "0769521223",

pages = "123--128",

booktitle = "Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition|Proc. Sixth IEEE Int. Conf. Autom. Face Gest. Recogn.",

publisher = "IEEE Computer Society ",

address = "United States",

note = "Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004 ; Conference date: 01-07-2004",

}

Bettinger, F & Cootes, TF 2004, A model of facial behaviour. in Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition|Proc. Sixth IEEE Int. Conf. Autom. Face Gest. Recogn.. IEEE Computer Society , pp. 123-128, Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004, Seoul, 1/07/04. https://doi.org/10.1109/AFGR.2004.1301519

TY - GEN

T1 - A model of facial behaviour

AU - Bettinger, F.

AU - Cootes, T. F.

PY - 2004

Y1 - 2004

N2 - We wish to model the way in which faces move in video sequences. We represent facial behaviour as a sequence of short actions. Each action is a sample from a statistical model representing the variability in the way it is performed. The ordering of actions is defined using a variable length Markov model. Action models and variable length Markov model are trained from a long (20000 frames) video sequence of a talking face. We propose a novel method of quantitatively evaluating the quality of the synthesis by measuring overlaps of parameter histograms. We apply this method to compare our technique with an alternative model that uses an autoregressive process.

AB - We wish to model the way in which faces move in video sequences. We represent facial behaviour as a sequence of short actions. Each action is a sample from a statistical model representing the variability in the way it is performed. The ordering of actions is defined using a variable length Markov model. Action models and variable length Markov model are trained from a long (20000 frames) video sequence of a talking face. We propose a novel method of quantitatively evaluating the quality of the synthesis by measuring overlaps of parameter histograms. We apply this method to compare our technique with an alternative model that uses an autoregressive process.

U2 - 10.1109/AFGR.2004.1301519

DO - 10.1109/AFGR.2004.1301519

M3 - Conference contribution

SN - 0769521223

SN - 9780769521220

SP - 123

EP - 128

BT - Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition|Proc. Sixth IEEE Int. Conf. Autom. Face Gest. Recogn.

PB - IEEE Computer Society

T2 - Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004

Y2 - 1 July 2004

ER -

A model of facial behaviour

Abstract

Conference

Access to Document

Fingerprint

Cite this