Automatic 3D Face Landmark Localization Based on 3D Vector Field Analysis

Syed Afaq A. Shah; Mohammed Bennamoun; Farid Boussaid

doi:10.1109/IVCNZ.2015.7761526

Automatic 3D Face Landmark Localization Based on 3D Vector Field Analysis

Syed Afaq A. Shah, Mohammed Bennamoun, Farid Boussaid

Research output: Chapter in Book/Conference paper › Conference paper › peer-review

1 Citation (Scopus)

Abstract

In applications such as 3D face synthesis and animation, a
prominent face landmark is required to enable 3D face normalization,
pose correction, 3D face recognition and reconstruction. Due to variations in facial expressions, automatic 3D face landmark localization remains a challenge. Nose tip is one of the salient landmarks in a human face. In this
paper, a novel nose tip localization technique is proposed. In the proposed approach, the rotation of the 3D vector field is analyzed for robust and efficient nose tip localization. The proposed technique has the following three characteristics: (1) it does not require any training; (2) it does not rely on
any particular model; (3) it is very efficient, requiring an average time of only 1.9s for nose tip detection. We tested the proposed technique on BU3DFE and Shrec’10 datasets. Experimental results show that the proposed technique is
robust to variations in facial expressions, achieving a 100% detection rate on these publicly available 3D face datasets.

Original language	English
Title of host publication	2015 International Conference on Image and Vision Computing New Zealand (IVCNZ)
Place of Publication	Auckland, NZ
Publisher	Wiley-IEEE Press
ISBN (Print)	9781509003570
DOIs	https://doi.org/10.1109/IVCNZ.2015.7761526
Publication status	Published - 2015
Event	30th International Conference on Image and Vision Computing, New Zealand (IVCNZ 2015) - Auckland, NZ, Auckland, New Zealand Duration: 23 Nov 2015 → 24 Nov 2015

Conference

Conference	30th International Conference on Image and Vision Computing, New Zealand (IVCNZ 2015)
Country/Territory	New Zealand
City	Auckland
Period	23/11/15 → 24/11/15

Access to Document

10.1109/IVCNZ.2015.7761526

Cite this

@inproceedings{ffe6f1a369194b2484a5e4761e819112,

title = "Automatic 3D Face Landmark Localization Based on 3D Vector Field Analysis",

abstract = "In applications such as 3D face synthesis and animation, aprominent face landmark is required to enable 3D face normalization,pose correction, 3D face recognition and reconstruction. Due to variations in facial expressions, automatic 3D face landmark localization remains a challenge. Nose tip is one of the salient landmarks in a human face. In thispaper, a novel nose tip localization technique is proposed. In the proposed approach, the rotation of the 3D vector field is analyzed for robust and efficient nose tip localization. The proposed technique has the following three characteristics: (1) it does not require any training; (2) it does not rely onany particular model; (3) it is very efficient, requiring an average time of only 1.9s for nose tip detection. We tested the proposed technique on BU3DFE and Shrec{\textquoteright}10 datasets. Experimental results show that the proposed technique isrobust to variations in facial expressions, achieving a 100% detection rate on these publicly available 3D face datasets.",

author = "Shah, {Syed Afaq A.} and Mohammed Bennamoun and Farid Boussaid",

year = "2015",

doi = "10.1109/IVCNZ.2015.7761526",

language = "English",

isbn = "9781509003570",

booktitle = "2015 International Conference on Image and Vision Computing New Zealand (IVCNZ)",

publisher = "Wiley-IEEE Press",

note = "30th International Conference on Image and Vision Computing, New Zealand (IVCNZ 2015) ; Conference date: 23-11-2015 Through 24-11-2015",

}

Shah, SAA , Bennamoun, M & Boussaid, F 2015, Automatic 3D Face Landmark Localization Based on 3D Vector Field Analysis. in 2015 International Conference on Image and Vision Computing New Zealand (IVCNZ). Wiley-IEEE Press, Auckland, NZ, 30th International Conference on Image and Vision Computing, New Zealand (IVCNZ 2015), Auckland, New Zealand, 23/11/15. https://doi.org/10.1109/IVCNZ.2015.7761526

TY - GEN

T1 - Automatic 3D Face Landmark Localization Based on 3D Vector Field Analysis

AU - Shah, Syed Afaq A.

AU - Bennamoun, Mohammed

AU - Boussaid, Farid

PY - 2015

Y1 - 2015

N2 - In applications such as 3D face synthesis and animation, aprominent face landmark is required to enable 3D face normalization,pose correction, 3D face recognition and reconstruction. Due to variations in facial expressions, automatic 3D face landmark localization remains a challenge. Nose tip is one of the salient landmarks in a human face. In thispaper, a novel nose tip localization technique is proposed. In the proposed approach, the rotation of the 3D vector field is analyzed for robust and efficient nose tip localization. The proposed technique has the following three characteristics: (1) it does not require any training; (2) it does not rely onany particular model; (3) it is very efficient, requiring an average time of only 1.9s for nose tip detection. We tested the proposed technique on BU3DFE and Shrec’10 datasets. Experimental results show that the proposed technique isrobust to variations in facial expressions, achieving a 100% detection rate on these publicly available 3D face datasets.

AB - In applications such as 3D face synthesis and animation, aprominent face landmark is required to enable 3D face normalization,pose correction, 3D face recognition and reconstruction. Due to variations in facial expressions, automatic 3D face landmark localization remains a challenge. Nose tip is one of the salient landmarks in a human face. In thispaper, a novel nose tip localization technique is proposed. In the proposed approach, the rotation of the 3D vector field is analyzed for robust and efficient nose tip localization. The proposed technique has the following three characteristics: (1) it does not require any training; (2) it does not rely onany particular model; (3) it is very efficient, requiring an average time of only 1.9s for nose tip detection. We tested the proposed technique on BU3DFE and Shrec’10 datasets. Experimental results show that the proposed technique isrobust to variations in facial expressions, achieving a 100% detection rate on these publicly available 3D face datasets.

U2 - 10.1109/IVCNZ.2015.7761526

DO - 10.1109/IVCNZ.2015.7761526

M3 - Conference paper

SN - 9781509003570

BT - 2015 International Conference on Image and Vision Computing New Zealand (IVCNZ)

PB - Wiley-IEEE Press

CY - Auckland, NZ

T2 - 30th International Conference on Image and Vision Computing, New Zealand (IVCNZ 2015)

Y2 - 23 November 2015 through 24 November 2015

ER -

Automatic 3D Face Landmark Localization Based on 3D Vector Field Analysis

Abstract

Conference

Access to Document

Fingerprint

Cite this