Quality-Aware Estimation of Facial Landmarks in Video Sequences

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

10 Citations (Scopus)
649 Downloads (Pure)

Abstract

Face alignment in video is a primitive step for facial
image analysis. The accuracy of the alignment greatly
depends on the quality of the face image in the video
frames and low quality faces are proven to cause
erroneous alignment. Thus, this paper proposes a system
for quality aware face alignment by using a Supervised
Decent Method (SDM) along with a motion based forward
extrapolation method. The proposed system first extracts
faces from video frames. Then, it employs a face quality
assessment technique to measure the face quality. If the
face quality is high, the proposed system uses SDM for
facial landmark detection. If the face quality is low the
proposed system corrects the facial landmarks that are
detected by SDM. Depending upon the face velocity in
consecutive video frames and face quality measure, two
algorithms are proposed for correction of landmarks in
low quality faces by using an extrapolation polynomial.
Experimental results illustrate the competency of the
proposed method while comparing with the state-of-theart
methods including an SDM-based method (from
CVPR-2013) and a very recent method (from CVPR-2014)
that uses parallel cascade of linear regression (Par-CLR).
Original languageEnglish
Title of host publicationIEEE Winter Conference on Applications of Computer Vision
Number of pages8
Place of PublicationUSA
PublisherIEEE Computer Society Press
Publication date6 Jan 2015
Pages678-685
Article number7045950
ISBN (Print)9781479966820
DOIs
Publication statusPublished - 6 Jan 2015
EventIEEE Winter Conference on Applications of Computer Vision (WACV) - Waikoloa Beach, Hawaii, United States
Duration: 6 Jan 20158 Jan 2015

Conference

ConferenceIEEE Winter Conference on Applications of Computer Vision (WACV)
CountryUnited States
CityWaikoloa Beach, Hawaii
Period06/01/201508/01/2015

Fingerprint

Linear regression
Extrapolation
Polynomials

Keywords

  • Facial Landmarks
  • Quality assessmnet
  • Tracking
  • Detection

Cite this

Haque, M. A., Nasrollahi, K., & Moeslund, T. B. (2015). Quality-Aware Estimation of Facial Landmarks in Video Sequences. In IEEE Winter Conference on Applications of Computer Vision (pp. 678-685). [7045950] USA: IEEE Computer Society Press. https://doi.org/10.1109/WACV.2015.96
Haque, Mohammad Ahsanul ; Nasrollahi, Kamal ; Moeslund, Thomas B. / Quality-Aware Estimation of Facial Landmarks in Video Sequences. IEEE Winter Conference on Applications of Computer Vision. USA : IEEE Computer Society Press, 2015. pp. 678-685
@inproceedings{f9dd6d1d05774c8ebae5367d91a0e508,
title = "Quality-Aware Estimation of Facial Landmarks in Video Sequences",
abstract = "Face alignment in video is a primitive step for facialimage analysis. The accuracy of the alignment greatlydepends on the quality of the face image in the videoframes and low quality faces are proven to causeerroneous alignment. Thus, this paper proposes a systemfor quality aware face alignment by using a SupervisedDecent Method (SDM) along with a motion based forwardextrapolation method. The proposed system first extractsfaces from video frames. Then, it employs a face qualityassessment technique to measure the face quality. If theface quality is high, the proposed system uses SDM forfacial landmark detection. If the face quality is low theproposed system corrects the facial landmarks that aredetected by SDM. Depending upon the face velocity inconsecutive video frames and face quality measure, twoalgorithms are proposed for correction of landmarks inlow quality faces by using an extrapolation polynomial.Experimental results illustrate the competency of theproposed method while comparing with the state-of-theartmethods including an SDM-based method (fromCVPR-2013) and a very recent method (from CVPR-2014)that uses parallel cascade of linear regression (Par-CLR).",
keywords = "Facial Landmarks, Quality assessmnet, Tracking, Detection",
author = "Haque, {Mohammad Ahsanul} and Kamal Nasrollahi and Moeslund, {Thomas B.}",
year = "2015",
month = "1",
day = "6",
doi = "10.1109/WACV.2015.96",
language = "English",
isbn = "9781479966820",
pages = "678--685",
booktitle = "IEEE Winter Conference on Applications of Computer Vision",
publisher = "IEEE Computer Society Press",
address = "United States",

}

Haque, MA, Nasrollahi, K & Moeslund, TB 2015, Quality-Aware Estimation of Facial Landmarks in Video Sequences. in IEEE Winter Conference on Applications of Computer Vision., 7045950, IEEE Computer Society Press, USA, pp. 678-685, IEEE Winter Conference on Applications of Computer Vision (WACV), Waikoloa Beach, Hawaii, United States, 06/01/2015. https://doi.org/10.1109/WACV.2015.96

Quality-Aware Estimation of Facial Landmarks in Video Sequences. / Haque, Mohammad Ahsanul; Nasrollahi, Kamal; Moeslund, Thomas B.

IEEE Winter Conference on Applications of Computer Vision. USA : IEEE Computer Society Press, 2015. p. 678-685 7045950.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Quality-Aware Estimation of Facial Landmarks in Video Sequences

AU - Haque, Mohammad Ahsanul

AU - Nasrollahi, Kamal

AU - Moeslund, Thomas B.

PY - 2015/1/6

Y1 - 2015/1/6

N2 - Face alignment in video is a primitive step for facialimage analysis. The accuracy of the alignment greatlydepends on the quality of the face image in the videoframes and low quality faces are proven to causeerroneous alignment. Thus, this paper proposes a systemfor quality aware face alignment by using a SupervisedDecent Method (SDM) along with a motion based forwardextrapolation method. The proposed system first extractsfaces from video frames. Then, it employs a face qualityassessment technique to measure the face quality. If theface quality is high, the proposed system uses SDM forfacial landmark detection. If the face quality is low theproposed system corrects the facial landmarks that aredetected by SDM. Depending upon the face velocity inconsecutive video frames and face quality measure, twoalgorithms are proposed for correction of landmarks inlow quality faces by using an extrapolation polynomial.Experimental results illustrate the competency of theproposed method while comparing with the state-of-theartmethods including an SDM-based method (fromCVPR-2013) and a very recent method (from CVPR-2014)that uses parallel cascade of linear regression (Par-CLR).

AB - Face alignment in video is a primitive step for facialimage analysis. The accuracy of the alignment greatlydepends on the quality of the face image in the videoframes and low quality faces are proven to causeerroneous alignment. Thus, this paper proposes a systemfor quality aware face alignment by using a SupervisedDecent Method (SDM) along with a motion based forwardextrapolation method. The proposed system first extractsfaces from video frames. Then, it employs a face qualityassessment technique to measure the face quality. If theface quality is high, the proposed system uses SDM forfacial landmark detection. If the face quality is low theproposed system corrects the facial landmarks that aredetected by SDM. Depending upon the face velocity inconsecutive video frames and face quality measure, twoalgorithms are proposed for correction of landmarks inlow quality faces by using an extrapolation polynomial.Experimental results illustrate the competency of theproposed method while comparing with the state-of-theartmethods including an SDM-based method (fromCVPR-2013) and a very recent method (from CVPR-2014)that uses parallel cascade of linear regression (Par-CLR).

KW - Facial Landmarks

KW - Quality assessmnet

KW - Tracking

KW - Detection

U2 - 10.1109/WACV.2015.96

DO - 10.1109/WACV.2015.96

M3 - Article in proceeding

SN - 9781479966820

SP - 678

EP - 685

BT - IEEE Winter Conference on Applications of Computer Vision

PB - IEEE Computer Society Press

CY - USA

ER -

Haque MA, Nasrollahi K, Moeslund TB. Quality-Aware Estimation of Facial Landmarks in Video Sequences. In IEEE Winter Conference on Applications of Computer Vision. USA: IEEE Computer Society Press. 2015. p. 678-685. 7045950 https://doi.org/10.1109/WACV.2015.96