A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs

Abstract

Recently, it has been shown that performance of a face recognition system depends on the quality of both face images participating in the recognition process: the reference and the test image. In the context of forensic face recognition, this observation has two implications: a) the quality of the trace (extracted from CCTV footage) constrains the performance achievable using a particular face recognition system; b) the quality of the suspect reference set (to which the trace is matched against) can be judiciously chosen to approach optimal recognition performance under such a constraint. Motivated by these recent findings, we propose a framework for forensic face recognition that is based on calibrating the recognition performance for the quality of pairs of images. The application of this framework to several mock-up forensic cases, created entirely from the MultiPIE dataset, shows that optimal recognition performance, under such a constraint, can be achieved by matching the quality (pose, illumination, and, imaging device) of the reference set to that of the trace. This improvement in recognition performance helps reduce the rate of misleading interpretation of the evidence.
Original languageUndefined
Place of PublicationEnschede
PublisherCentre for Telematics and Information Technology (CTIT)
StatePublished - Sep 2012

Publication series

NameCTIT Technical Report Series
PublisherCentre for Telematics and Information Technology, University of Twente
No.TR-CTIT-12-20
ISSN (Print)1381-3625

Fingerprint

Face recognition
Closed circuit television systems
Lighting
Imaging techniques

Keywords

  • IR-84354
  • METIS-289688
  • SCS-Safety
  • EWI-22237

Cite this

Dutta, A., Veldhuis, R. N. J., Spreeuwers, L. J., Meuwly, D., & Meuwly, D. (2012). A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs. (CTIT Technical Report Series; No. TR-CTIT-12-20). Enschede: Centre for Telematics and Information Technology (CTIT).

Dutta, A.; Veldhuis, Raymond N.J.; Spreeuwers, Lieuwe Jan; Meuwly, Didier; Meuwly, Didier / A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs.

Enschede : Centre for Telematics and Information Technology (CTIT), 2012. (CTIT Technical Report Series; No. TR-CTIT-12-20).

Research output: ProfessionalReport

@book{a173de55874044beaaa1da7d0e3cabea,
title = "A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs",
abstract = "Recently, it has been shown that performance of a face recognition system depends on the quality of both face images participating in the recognition process: the reference and the test image. In the context of forensic face recognition, this observation has two implications: a) the quality of the trace (extracted from CCTV footage) constrains the performance achievable using a particular face recognition system; b) the quality of the suspect reference set (to which the trace is matched against) can be judiciously chosen to approach optimal recognition performance under such a constraint. Motivated by these recent findings, we propose a framework for forensic face recognition that is based on calibrating the recognition performance for the quality of pairs of images. The application of this framework to several mock-up forensic cases, created entirely from the MultiPIE dataset, shows that optimal recognition performance, under such a constraint, can be achieved by matching the quality (pose, illumination, and, imaging device) of the reference set to that of the trace. This improvement in recognition performance helps reduce the rate of misleading interpretation of the evidence.",
keywords = "IR-84354, METIS-289688, SCS-Safety, EWI-22237",
author = "A. Dutta and Veldhuis, {Raymond N.J.} and Spreeuwers, {Lieuwe Jan} and Didier Meuwly and Didier Meuwly",
year = "2012",
month = "9",
series = "CTIT Technical Report Series",
publisher = "Centre for Telematics and Information Technology (CTIT)",
number = "TR-CTIT-12-20",
address = "Netherlands",

}

Dutta, A, Veldhuis, RNJ, Spreeuwers, LJ, Meuwly, D & Meuwly, D 2012, A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs. CTIT Technical Report Series, no. TR-CTIT-12-20, Centre for Telematics and Information Technology (CTIT), Enschede.

A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs. / Dutta, A.; Veldhuis, Raymond N.J.; Spreeuwers, Lieuwe Jan; Meuwly, Didier; Meuwly, Didier.

Enschede : Centre for Telematics and Information Technology (CTIT), 2012. (CTIT Technical Report Series; No. TR-CTIT-12-20).

Research output: ProfessionalReport

TY - BOOK

T1 - A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs

AU - Dutta,A.

AU - Veldhuis,Raymond N.J.

AU - Spreeuwers,Lieuwe Jan

AU - Meuwly,Didier

AU - Meuwly,Didier

PY - 2012/9

Y1 - 2012/9

N2 - Recently, it has been shown that performance of a face recognition system depends on the quality of both face images participating in the recognition process: the reference and the test image. In the context of forensic face recognition, this observation has two implications: a) the quality of the trace (extracted from CCTV footage) constrains the performance achievable using a particular face recognition system; b) the quality of the suspect reference set (to which the trace is matched against) can be judiciously chosen to approach optimal recognition performance under such a constraint. Motivated by these recent findings, we propose a framework for forensic face recognition that is based on calibrating the recognition performance for the quality of pairs of images. The application of this framework to several mock-up forensic cases, created entirely from the MultiPIE dataset, shows that optimal recognition performance, under such a constraint, can be achieved by matching the quality (pose, illumination, and, imaging device) of the reference set to that of the trace. This improvement in recognition performance helps reduce the rate of misleading interpretation of the evidence.

AB - Recently, it has been shown that performance of a face recognition system depends on the quality of both face images participating in the recognition process: the reference and the test image. In the context of forensic face recognition, this observation has two implications: a) the quality of the trace (extracted from CCTV footage) constrains the performance achievable using a particular face recognition system; b) the quality of the suspect reference set (to which the trace is matched against) can be judiciously chosen to approach optimal recognition performance under such a constraint. Motivated by these recent findings, we propose a framework for forensic face recognition that is based on calibrating the recognition performance for the quality of pairs of images. The application of this framework to several mock-up forensic cases, created entirely from the MultiPIE dataset, shows that optimal recognition performance, under such a constraint, can be achieved by matching the quality (pose, illumination, and, imaging device) of the reference set to that of the trace. This improvement in recognition performance helps reduce the rate of misleading interpretation of the evidence.

KW - IR-84354

KW - METIS-289688

KW - SCS-Safety

KW - EWI-22237

M3 - Report

T3 - CTIT Technical Report Series

BT - A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs

PB - Centre for Telematics and Information Technology (CTIT)

ER -

Dutta A, Veldhuis RNJ, Spreeuwers LJ, Meuwly D, Meuwly D. A framework for forensic face recognition based on recognition performance calibrated for the quality of image pairs. Enschede: Centre for Telematics and Information Technology (CTIT), 2012. (CTIT Technical Report Series; TR-CTIT-12-20).