Realistic 3D facial animation parameters from mirror-reflected multi-view video

I-Chen Lin; Jeng Sheng Yeh; Ming Ouhyoung

doi:10.1109/CA.2001.982371

Realistic 3D facial animation parameters from mirror-reflected multi-view video

I-Chen Lin^*, Jeng Sheng Yeh, Ming Ouhyoung

^*Corresponding author for this work

Institute of Multimedia Engineering

Research output: Contribution to conference › Paper › peer-review

11 Scopus citations

Abstract

In this paper, a robust, accurate and inexpensive approach to estimate 3D facial motion from multi-view video is proposed, where two mirrors located near one's cheeks can reflect the side views of markers on one's face. Nice properties of mirrored images are utilized to simplify the proposed tracking algorithm significantly, while a Kalman filter is employed to reduce the noise and to predict the occluded markers positions. More than 50 markers on one's face are continuously tracked at 30 frames per second. The estimated 3D facial motion data has been practically applied to our facial animation system. In addition, the dataset of facial motion can also be applied to the analysis of co-articulation effects, facial expressions, and audio-visual hybrid recognition system.

Original language	English
Pages	2-11
Number of pages	10
DOIs	https://doi.org/10.1109/CA.2001.982371
State	Published - 1 Dec 2001
Event	14th Conference on Computer Animation - Seoul, Japan Duration: 7 Nov 2001 → 8 Nov 2001

Conference

Conference	14th Conference on Computer Animation
Country	Japan
City	Seoul
Period	7/11/01 → 8/11/01

Access to Document

10.1109/CA.2001.982371

Fingerprint Dive into the research topics of 'Realistic 3D facial animation parameters from mirror-reflected multi-view video'. Together they form a unique fingerprint.

Cite this

@conference{80a47990b5ae4d8ba5251bcd573247b0,

title = "Realistic 3D facial animation parameters from mirror-reflected multi-view video",

abstract = "In this paper, a robust, accurate and inexpensive approach to estimate 3D facial motion from multi-view video is proposed, where two mirrors located near one's cheeks can reflect the side views of markers on one's face. Nice properties of mirrored images are utilized to simplify the proposed tracking algorithm significantly, while a Kalman filter is employed to reduce the noise and to predict the occluded markers positions. More than 50 markers on one's face are continuously tracked at 30 frames per second. The estimated 3D facial motion data has been practically applied to our facial animation system. In addition, the dataset of facial motion can also be applied to the analysis of co-articulation effects, facial expressions, and audio-visual hybrid recognition system.",

author = "I-Chen Lin and Yeh, {Jeng Sheng} and Ming Ouhyoung",

year = "2001",

month = dec,

day = "1",

doi = "10.1109/CA.2001.982371",

language = "English",

pages = "2--11",

note = "null ; Conference date: 07-11-2001 Through 08-11-2001",

}

TY - CONF

T1 - Realistic 3D facial animation parameters from mirror-reflected multi-view video

AU - Lin, I-Chen

AU - Yeh, Jeng Sheng

AU - Ouhyoung, Ming

PY - 2001/12/1

Y1 - 2001/12/1

N2 - In this paper, a robust, accurate and inexpensive approach to estimate 3D facial motion from multi-view video is proposed, where two mirrors located near one's cheeks can reflect the side views of markers on one's face. Nice properties of mirrored images are utilized to simplify the proposed tracking algorithm significantly, while a Kalman filter is employed to reduce the noise and to predict the occluded markers positions. More than 50 markers on one's face are continuously tracked at 30 frames per second. The estimated 3D facial motion data has been practically applied to our facial animation system. In addition, the dataset of facial motion can also be applied to the analysis of co-articulation effects, facial expressions, and audio-visual hybrid recognition system.

AB - In this paper, a robust, accurate and inexpensive approach to estimate 3D facial motion from multi-view video is proposed, where two mirrors located near one's cheeks can reflect the side views of markers on one's face. Nice properties of mirrored images are utilized to simplify the proposed tracking algorithm significantly, while a Kalman filter is employed to reduce the noise and to predict the occluded markers positions. More than 50 markers on one's face are continuously tracked at 30 frames per second. The estimated 3D facial motion data has been practically applied to our facial animation system. In addition, the dataset of facial motion can also be applied to the analysis of co-articulation effects, facial expressions, and audio-visual hybrid recognition system.

UR - http://www.scopus.com/inward/record.url?scp=0035556994&partnerID=8YFLogxK

U2 - 10.1109/CA.2001.982371

DO - 10.1109/CA.2001.982371

M3 - Paper

AN - SCOPUS:0035556994

SP - 2

EP - 11

Y2 - 7 November 2001 through 8 November 2001

ER -