Face identification in videos from mobile cameras

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

35 Downloads (Pure)

Abstract

It is still challenging to recognize faces reliably in videos from mobile camera, although mature automatic face recognition technology for still images has been available for quite some time. Suppose we want to be alerted when suspects appear in the recording of a police Body-Cam, even a good face matcher on still images would give many false alarms due to the uncontrolled conditions. This paper presents an approach to identify faces in videos from mobile cameras. A commercial face matcher FaceVACS is used to process the face recognition frame by frame. On a video of certain length, in order to suppress the false alarms, we propose to count the recognized identities and set thresholds to the counts, as well as to the matching scores for still-image face recognition. In this way, the facial information of a single subject over time is exploited without implementing face tracking, which is complicated and more difficult for low-quality unconstrained videos. For experiments, videos are recorded by two type of mobile cameras, which provide different video qualities. The results demonstrate the efficiency of our proposed approach.
Original languageUndefined
Title of host publicationProceedings of Netherlands Conference on Computer Vision, NCCV 2014
Place of PublicationDelft
PublisherASCI
Pages1-4
Number of pages4
ISBN (Print)not assigned
Publication statusPublished - 24 Apr 2014

Publication series

Name
PublisherASCI

Keywords

  • EWI-25069
  • SCS-Safety
  • Face Recognition
  • METIS-306027
  • Biometrics
  • Mobile camera
  • IR-91811

Cite this

Mu, M., Spreeuwers, L. J., & Veldhuis, R. N. J. (2014). Face identification in videos from mobile cameras. In Proceedings of Netherlands Conference on Computer Vision, NCCV 2014 (pp. 1-4). Delft: ASCI.
Mu, Meiru ; Spreeuwers, Lieuwe Jan ; Veldhuis, Raymond N.J. / Face identification in videos from mobile cameras. Proceedings of Netherlands Conference on Computer Vision, NCCV 2014. Delft : ASCI, 2014. pp. 1-4
@inproceedings{f9c0de71ff9b4016b9eadef9f9de17da,
title = "Face identification in videos from mobile cameras",
abstract = "It is still challenging to recognize faces reliably in videos from mobile camera, although mature automatic face recognition technology for still images has been available for quite some time. Suppose we want to be alerted when suspects appear in the recording of a police Body-Cam, even a good face matcher on still images would give many false alarms due to the uncontrolled conditions. This paper presents an approach to identify faces in videos from mobile cameras. A commercial face matcher FaceVACS is used to process the face recognition frame by frame. On a video of certain length, in order to suppress the false alarms, we propose to count the recognized identities and set thresholds to the counts, as well as to the matching scores for still-image face recognition. In this way, the facial information of a single subject over time is exploited without implementing face tracking, which is complicated and more difficult for low-quality unconstrained videos. For experiments, videos are recorded by two type of mobile cameras, which provide different video qualities. The results demonstrate the efficiency of our proposed approach.",
keywords = "EWI-25069, SCS-Safety, Face Recognition, METIS-306027, Biometrics, Mobile camera, IR-91811",
author = "Meiru Mu and Spreeuwers, {Lieuwe Jan} and Veldhuis, {Raymond N.J.}",
note = "eemcs-eprint-25069",
year = "2014",
month = "4",
day = "24",
language = "Undefined",
isbn = "not assigned",
publisher = "ASCI",
pages = "1--4",
booktitle = "Proceedings of Netherlands Conference on Computer Vision, NCCV 2014",

}

Mu, M, Spreeuwers, LJ & Veldhuis, RNJ 2014, Face identification in videos from mobile cameras. in Proceedings of Netherlands Conference on Computer Vision, NCCV 2014. ASCI, Delft, pp. 1-4.

Face identification in videos from mobile cameras. / Mu, Meiru; Spreeuwers, Lieuwe Jan; Veldhuis, Raymond N.J.

Proceedings of Netherlands Conference on Computer Vision, NCCV 2014. Delft : ASCI, 2014. p. 1-4.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Face identification in videos from mobile cameras

AU - Mu, Meiru

AU - Spreeuwers, Lieuwe Jan

AU - Veldhuis, Raymond N.J.

N1 - eemcs-eprint-25069

PY - 2014/4/24

Y1 - 2014/4/24

N2 - It is still challenging to recognize faces reliably in videos from mobile camera, although mature automatic face recognition technology for still images has been available for quite some time. Suppose we want to be alerted when suspects appear in the recording of a police Body-Cam, even a good face matcher on still images would give many false alarms due to the uncontrolled conditions. This paper presents an approach to identify faces in videos from mobile cameras. A commercial face matcher FaceVACS is used to process the face recognition frame by frame. On a video of certain length, in order to suppress the false alarms, we propose to count the recognized identities and set thresholds to the counts, as well as to the matching scores for still-image face recognition. In this way, the facial information of a single subject over time is exploited without implementing face tracking, which is complicated and more difficult for low-quality unconstrained videos. For experiments, videos are recorded by two type of mobile cameras, which provide different video qualities. The results demonstrate the efficiency of our proposed approach.

AB - It is still challenging to recognize faces reliably in videos from mobile camera, although mature automatic face recognition technology for still images has been available for quite some time. Suppose we want to be alerted when suspects appear in the recording of a police Body-Cam, even a good face matcher on still images would give many false alarms due to the uncontrolled conditions. This paper presents an approach to identify faces in videos from mobile cameras. A commercial face matcher FaceVACS is used to process the face recognition frame by frame. On a video of certain length, in order to suppress the false alarms, we propose to count the recognized identities and set thresholds to the counts, as well as to the matching scores for still-image face recognition. In this way, the facial information of a single subject over time is exploited without implementing face tracking, which is complicated and more difficult for low-quality unconstrained videos. For experiments, videos are recorded by two type of mobile cameras, which provide different video qualities. The results demonstrate the efficiency of our proposed approach.

KW - EWI-25069

KW - SCS-Safety

KW - Face Recognition

KW - METIS-306027

KW - Biometrics

KW - Mobile camera

KW - IR-91811

M3 - Conference contribution

SN - not assigned

SP - 1

EP - 4

BT - Proceedings of Netherlands Conference on Computer Vision, NCCV 2014

PB - ASCI

CY - Delft

ER -

Mu M, Spreeuwers LJ, Veldhuis RNJ. Face identification in videos from mobile cameras. In Proceedings of Netherlands Conference on Computer Vision, NCCV 2014. Delft: ASCI. 2014. p. 1-4