Estimation of Face Parameters using Correlation Analysis and a Topology Preserving Prior

Stella  Grasshof; Hanno Ackermann; Jörn Ostermann

doi:10.1109/MVA.2015.7153259

Estimation of Face Parameters using Correlation Analysis and a Topology Preserving Prior

Stella Grasshof
, Hanno Ackermann
, Jörn Ostermann

Leibniz University Hannover

Research output: Conference Article in Proceeding or Book/Report chapter › Article in proceedings › Research › peer-review

Abstract

Candide-3 is a well-known model, used to represent triangular meshes of human faces. It is common to only estimate 17 to 21 of the 79 model parameters. We show that these are insufficient to fit model vertices to facial feature points with low error and if more parameters are estimated, the model mesh deforms to unnatural configurations. To overcome this problem, we propose a novel solution: Given facial feature points, we propose to estimate the model parameters in subsets in which they are uncorrelated. Additionally we present a term to penalize topologically incorrect triangular mesh configurations. As a result the average mean squared error between facial feature points and model vertices is reduced by 90%, while face topology is preserved.

Original language	English
Title of host publication	2015 14th IAPR International Conference on Machine Vision Applications (MVA)
Publisher	IEEE
Publication date	2015
Pages	584-587
ISBN (Electronic)	978-4-9011-2214-6
DOIs	https://doi.org/10.1109/MVA.2015.7153259
Publication status	Published - 2015
Externally published	Yes
Event	The Fourteenth IAPR International Conference on Machine Vision Applications - Tokyo, Japan Duration: 18 May 2015 → 22 May 2015 https://www.mva-org.jp/mva2015/

Conference

Conference	The Fourteenth IAPR International Conference on Machine Vision Applications
Country/Territory	Japan
City	Tokyo
Period	18/05/2015 → 22/05/2015
Internet address	https://www.mva-org.jp/mva2015/

Series	Proceedings of the IAPR International Conference on Machine Vision Applications (MVA)

Keywords

Face
Shape
Facial features
Solid modeling
Three-dimensional displays
Estimation
Topology

Access to Document

10.1109/MVA.2015.7153259Licence: Unspecified

Cite this

@inproceedings{38f190078fa04bcdb924ac44114f7c14,

title = "Estimation of Face Parameters using Correlation Analysis and a Topology Preserving Prior",

abstract = "Candide-3 is a well-known model, used to represent triangular meshes of human faces. It is common to only estimate 17 to 21 of the 79 model parameters. We show that these are insufficient to fit model vertices to facial feature points with low error and if more parameters are estimated, the model mesh deforms to unnatural configurations. To overcome this problem, we propose a novel solution: Given facial feature points, we propose to estimate the model parameters in subsets in which they are uncorrelated. Additionally we present a term to penalize topologically incorrect triangular mesh configurations. As a result the average mean squared error between facial feature points and model vertices is reduced by 90\%, while face topology is preserved.",

keywords = "Face, Shape, Facial features, Solid modeling, Three-dimensional displays, Estimation, Topology",

author = "Stella Grasshof and Hanno Ackermann and J{\"o}rn Ostermann",

year = "2015",

doi = "10.1109/MVA.2015.7153259",

language = "English",

series = "Proceedings of the IAPR International Conference on Machine Vision Applications (MVA)",

publisher = "IEEE",

pages = "584--587",

booktitle = "2015 14th IAPR International Conference on Machine Vision Applications (MVA)",

address = "United States",

note = "The Fourteenth IAPR International Conference on Machine Vision Applications, MVA2015 ; Conference date: 18-05-2015 Through 22-05-2015",

url = "https://www.mva-org.jp/mva2015/",

}

Grasshof, S, Ackermann, H & Ostermann, J 2015, Estimation of Face Parameters using Correlation Analysis and a Topology Preserving Prior. in 2015 14th IAPR International Conference on Machine Vision Applications (MVA). IEEE, Proceedings of the IAPR International Conference on Machine Vision Applications (MVA), pp. 584-587, The Fourteenth IAPR International Conference on Machine Vision Applications, Tokyo, Japan, 18/05/2015. https://doi.org/10.1109/MVA.2015.7153259

Estimation of Face Parameters using Correlation Analysis and a Topology Preserving Prior. / Grasshof, Stella ; Ackermann, Hanno; Ostermann, Jörn.
2015 14th IAPR International Conference on Machine Vision Applications (MVA). IEEE, 2015. p. 584-587 (Proceedings of the IAPR International Conference on Machine Vision Applications (MVA)).

Research output: Conference Article in Proceeding or Book/Report chapter › Article in proceedings › Research › peer-review

TY - GEN

T1 - Estimation of Face Parameters using Correlation Analysis and a Topology Preserving Prior

AU - Grasshof, Stella

AU - Ackermann, Hanno

AU - Ostermann, Jörn

PY - 2015

Y1 - 2015

N2 - Candide-3 is a well-known model, used to represent triangular meshes of human faces. It is common to only estimate 17 to 21 of the 79 model parameters. We show that these are insufficient to fit model vertices to facial feature points with low error and if more parameters are estimated, the model mesh deforms to unnatural configurations. To overcome this problem, we propose a novel solution: Given facial feature points, we propose to estimate the model parameters in subsets in which they are uncorrelated. Additionally we present a term to penalize topologically incorrect triangular mesh configurations. As a result the average mean squared error between facial feature points and model vertices is reduced by 90%, while face topology is preserved.

AB - Candide-3 is a well-known model, used to represent triangular meshes of human faces. It is common to only estimate 17 to 21 of the 79 model parameters. We show that these are insufficient to fit model vertices to facial feature points with low error and if more parameters are estimated, the model mesh deforms to unnatural configurations. To overcome this problem, we propose a novel solution: Given facial feature points, we propose to estimate the model parameters in subsets in which they are uncorrelated. Additionally we present a term to penalize topologically incorrect triangular mesh configurations. As a result the average mean squared error between facial feature points and model vertices is reduced by 90%, while face topology is preserved.

KW - Face

KW - Shape

KW - Facial features

KW - Solid modeling

KW - Three-dimensional displays

KW - Estimation

KW - Topology

U2 - 10.1109/MVA.2015.7153259

DO - 10.1109/MVA.2015.7153259

M3 - Article in proceedings

T3 - Proceedings of the IAPR International Conference on Machine Vision Applications (MVA)

SP - 584

EP - 587

BT - 2015 14th IAPR International Conference on Machine Vision Applications (MVA)

PB - IEEE

T2 - The Fourteenth IAPR International Conference on Machine Vision Applications

Y2 - 18 May 2015 through 22 May 2015

ER -

Estimation of Face Parameters using Correlation Analysis and a Topology Preserving Prior

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this