Contextual weighting of patches for local matching in still-to-video face recognition

Ibtihel Amara, Eric Granger, Abdenour Hadid

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Still-to-video face recognition (FR) systems for watchlist screening seek to recognize individuals of interest given faces captured over a network of video surveillance cameras. Screening faces against a watchlist is a challenging application because only a limited number of reference stills is available per individual during enrollment, and the appearance of face captures in videos changes from camera to camera, due to variations in illumination, pose, blur, scale, expression and occlusion. In order to improve the robustness of FR systems, several local matching techniques have been proposed that rely on static or dynamic weighting of patches. However, these approaches are not suitable for watchlist screening applications where the capturing conditions vary significantly over different camera fields of view (FoV). In this paper, a new dynamic weighting technique is proposed for weighting facial patches based on video data collected a priori from the specific operational domain (camera FoV) and on image quality assessment. Results obtained on videos from the Chokepoint dataset indicate that the proposed approach can significantly outperform the reference local matching methods because patch weights tend to grow for discriminant facial regions.

Original languageEnglish
Title of host publicationProceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages756-763
Number of pages8
ISBN (Electronic)9781538623350
DOIs
Publication statusPublished - 5 Jun 2018
Event13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018 - Xi'an, China
Duration: 15 May 201819 May 2018

Other

Other13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018
CountryChina
CityXi'an
Period15/5/1819/5/18

Fingerprint

Face recognition
Face Recognition
Weighting
Patch
Camera
Cameras
Screening
Field of View
Face
Image Quality Assessment
Video Surveillance
Discriminant
Occlusion
Image quality
Illumination
Lighting
Vary
Tend
Robustness

Keywords

  • Face Recognition
  • Video surveillance
  • Watchlist screening

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Control and Optimization

Cite this

Amara, I., Granger, E., & Hadid, A. (2018). Contextual weighting of patches for local matching in still-to-video face recognition. In Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018 (pp. 756-763). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/FG.2018.00119

Contextual weighting of patches for local matching in still-to-video face recognition. / Amara, Ibtihel; Granger, Eric; Hadid, Abdenour.

Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 756-763.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Amara, I, Granger, E & Hadid, A 2018, Contextual weighting of patches for local matching in still-to-video face recognition. in Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. Institute of Electrical and Electronics Engineers Inc., pp. 756-763, 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018, Xi'an, China, 15/5/18. https://doi.org/10.1109/FG.2018.00119
Amara I, Granger E, Hadid A. Contextual weighting of patches for local matching in still-to-video face recognition. In Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. Institute of Electrical and Electronics Engineers Inc. 2018. p. 756-763 https://doi.org/10.1109/FG.2018.00119
Amara, Ibtihel ; Granger, Eric ; Hadid, Abdenour. / Contextual weighting of patches for local matching in still-to-video face recognition. Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 756-763
@inproceedings{88d73a48ce59411a90dfda6d12a04a06,
title = "Contextual weighting of patches for local matching in still-to-video face recognition",
abstract = "Still-to-video face recognition (FR) systems for watchlist screening seek to recognize individuals of interest given faces captured over a network of video surveillance cameras. Screening faces against a watchlist is a challenging application because only a limited number of reference stills is available per individual during enrollment, and the appearance of face captures in videos changes from camera to camera, due to variations in illumination, pose, blur, scale, expression and occlusion. In order to improve the robustness of FR systems, several local matching techniques have been proposed that rely on static or dynamic weighting of patches. However, these approaches are not suitable for watchlist screening applications where the capturing conditions vary significantly over different camera fields of view (FoV). In this paper, a new dynamic weighting technique is proposed for weighting facial patches based on video data collected a priori from the specific operational domain (camera FoV) and on image quality assessment. Results obtained on videos from the Chokepoint dataset indicate that the proposed approach can significantly outperform the reference local matching methods because patch weights tend to grow for discriminant facial regions.",
keywords = "Face Recognition, Video surveillance, Watchlist screening",
author = "Ibtihel Amara and Eric Granger and Abdenour Hadid",
year = "2018",
month = "6",
day = "5",
doi = "10.1109/FG.2018.00119",
language = "English",
pages = "756--763",
booktitle = "Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Contextual weighting of patches for local matching in still-to-video face recognition

AU - Amara, Ibtihel

AU - Granger, Eric

AU - Hadid, Abdenour

PY - 2018/6/5

Y1 - 2018/6/5

N2 - Still-to-video face recognition (FR) systems for watchlist screening seek to recognize individuals of interest given faces captured over a network of video surveillance cameras. Screening faces against a watchlist is a challenging application because only a limited number of reference stills is available per individual during enrollment, and the appearance of face captures in videos changes from camera to camera, due to variations in illumination, pose, blur, scale, expression and occlusion. In order to improve the robustness of FR systems, several local matching techniques have been proposed that rely on static or dynamic weighting of patches. However, these approaches are not suitable for watchlist screening applications where the capturing conditions vary significantly over different camera fields of view (FoV). In this paper, a new dynamic weighting technique is proposed for weighting facial patches based on video data collected a priori from the specific operational domain (camera FoV) and on image quality assessment. Results obtained on videos from the Chokepoint dataset indicate that the proposed approach can significantly outperform the reference local matching methods because patch weights tend to grow for discriminant facial regions.

AB - Still-to-video face recognition (FR) systems for watchlist screening seek to recognize individuals of interest given faces captured over a network of video surveillance cameras. Screening faces against a watchlist is a challenging application because only a limited number of reference stills is available per individual during enrollment, and the appearance of face captures in videos changes from camera to camera, due to variations in illumination, pose, blur, scale, expression and occlusion. In order to improve the robustness of FR systems, several local matching techniques have been proposed that rely on static or dynamic weighting of patches. However, these approaches are not suitable for watchlist screening applications where the capturing conditions vary significantly over different camera fields of view (FoV). In this paper, a new dynamic weighting technique is proposed for weighting facial patches based on video data collected a priori from the specific operational domain (camera FoV) and on image quality assessment. Results obtained on videos from the Chokepoint dataset indicate that the proposed approach can significantly outperform the reference local matching methods because patch weights tend to grow for discriminant facial regions.

KW - Face Recognition

KW - Video surveillance

KW - Watchlist screening

UR - http://www.scopus.com/inward/record.url?scp=85049399056&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049399056&partnerID=8YFLogxK

U2 - 10.1109/FG.2018.00119

DO - 10.1109/FG.2018.00119

M3 - Conference contribution

AN - SCOPUS:85049399056

SP - 756

EP - 763

BT - Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -