3D Face Geometry Recovery Via Group-wise Optical Flow

Hui Fang, Nick Costen, David Cristinacce, John Darby

Research output: Contribution to conferencePoster

4 Citations (Scopus)

Abstract

We describe an algorithm for automatically finding correspondences from face video sequences. This method is useful to many applications such as face tracking, face modeling and 3D face recovery. Given a sequence of images, the face feature points are tracked by a model-constraint optical flow algorithm. By employing a Minimum Description Length (MDL) point-refinement framework, the drift-off error caused by the optical flow algorithm can be reduced and the correspondences can be matched robustly by optimizing the statistical model. As a result, the face is able to be tracked precisely. Furthermore, it offers a new method of building an appearance model automatically. The objective root mean square error (RMSE) is used to prove the efficiency of the algorithm. At the same time, the performance is evaluated subjectively by generating 3D face models based upon it.
Original languageEnglish
Pages1-8
DOIs
Publication statusPublished - 2008
EventIEEE International Conference on Facial and gesture recognition - Amsterdam, Netherlands
Duration: 1 Jan 2008 → …

Conference

ConferenceIEEE International Conference on Facial and gesture recognition
CountryNetherlands
CityAmsterdam
Period1/01/08 → …

Fingerprint

Optical flows
Recovery
Geometry
Mean square error

Cite this

Fang, H., Costen, N., Cristinacce, D., & Darby, J. (2008). 3D Face Geometry Recovery Via Group-wise Optical Flow. 1-8. Poster session presented at IEEE International Conference on Facial and gesture recognition, Amsterdam, Netherlands. https://doi.org/10.1109/AFGR.2008.4813303
Fang, Hui ; Costen, Nick ; Cristinacce, David ; Darby, John. / 3D Face Geometry Recovery Via Group-wise Optical Flow. Poster session presented at IEEE International Conference on Facial and gesture recognition, Amsterdam, Netherlands.
@conference{5a71b35e5daf421eb43e0b2557ad68bc,
title = "3D Face Geometry Recovery Via Group-wise Optical Flow",
abstract = "We describe an algorithm for automatically finding correspondences from face video sequences. This method is useful to many applications such as face tracking, face modeling and 3D face recovery. Given a sequence of images, the face feature points are tracked by a model-constraint optical flow algorithm. By employing a Minimum Description Length (MDL) point-refinement framework, the drift-off error caused by the optical flow algorithm can be reduced and the correspondences can be matched robustly by optimizing the statistical model. As a result, the face is able to be tracked precisely. Furthermore, it offers a new method of building an appearance model automatically. The objective root mean square error (RMSE) is used to prove the efficiency of the algorithm. At the same time, the performance is evaluated subjectively by generating 3D face models based upon it.",
author = "Hui Fang and Nick Costen and David Cristinacce and John Darby",
year = "2008",
doi = "10.1109/AFGR.2008.4813303",
language = "English",
pages = "1--8",
note = "IEEE International Conference on Facial and gesture recognition ; Conference date: 01-01-2008",

}

Fang, H, Costen, N, Cristinacce, D & Darby, J 2008, '3D Face Geometry Recovery Via Group-wise Optical Flow' IEEE International Conference on Facial and gesture recognition, Amsterdam, Netherlands, 1/01/08, pp. 1-8. https://doi.org/10.1109/AFGR.2008.4813303

3D Face Geometry Recovery Via Group-wise Optical Flow. / Fang, Hui; Costen, Nick; Cristinacce, David; Darby, John.

2008. 1-8 Poster session presented at IEEE International Conference on Facial and gesture recognition, Amsterdam, Netherlands.

Research output: Contribution to conferencePoster

TY - CONF

T1 - 3D Face Geometry Recovery Via Group-wise Optical Flow

AU - Fang, Hui

AU - Costen, Nick

AU - Cristinacce, David

AU - Darby, John

PY - 2008

Y1 - 2008

N2 - We describe an algorithm for automatically finding correspondences from face video sequences. This method is useful to many applications such as face tracking, face modeling and 3D face recovery. Given a sequence of images, the face feature points are tracked by a model-constraint optical flow algorithm. By employing a Minimum Description Length (MDL) point-refinement framework, the drift-off error caused by the optical flow algorithm can be reduced and the correspondences can be matched robustly by optimizing the statistical model. As a result, the face is able to be tracked precisely. Furthermore, it offers a new method of building an appearance model automatically. The objective root mean square error (RMSE) is used to prove the efficiency of the algorithm. At the same time, the performance is evaluated subjectively by generating 3D face models based upon it.

AB - We describe an algorithm for automatically finding correspondences from face video sequences. This method is useful to many applications such as face tracking, face modeling and 3D face recovery. Given a sequence of images, the face feature points are tracked by a model-constraint optical flow algorithm. By employing a Minimum Description Length (MDL) point-refinement framework, the drift-off error caused by the optical flow algorithm can be reduced and the correspondences can be matched robustly by optimizing the statistical model. As a result, the face is able to be tracked precisely. Furthermore, it offers a new method of building an appearance model automatically. The objective root mean square error (RMSE) is used to prove the efficiency of the algorithm. At the same time, the performance is evaluated subjectively by generating 3D face models based upon it.

U2 - 10.1109/AFGR.2008.4813303

DO - 10.1109/AFGR.2008.4813303

M3 - Poster

SP - 1

EP - 8

ER -

Fang H, Costen N, Cristinacce D, Darby J. 3D Face Geometry Recovery Via Group-wise Optical Flow. 2008. Poster session presented at IEEE International Conference on Facial and gesture recognition, Amsterdam, Netherlands. https://doi.org/10.1109/AFGR.2008.4813303