Discriminating Power of FISWG Characteristic Descriptors Under Different Forensic Use Cases

Christopher Gerard Zeinstra; Raymond N.J. Veldhuis; Lieuwe Jan Spreeuwers

doi:10.1109/BIOSIG.2016.7736919

Discriminating Power of FISWG Characteristic Descriptors Under Different Forensic Use Cases

Christopher Gerard Zeinstra
, Raymond N.J. Veldhuis
, Lieuwe Jan Spreeuwers

Datamanagement & Biometrics

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

3 Citations (Scopus)

29 Downloads (Pure)

Abstract

FISWG characteristic descriptors are facial features that can be used for evidence evaluation during forensic case work. In this paper we investigate the discriminating power of a biometric system that uses these characteristic descriptors as features under different forensic use cases. We show that in every forensic use case we can find characteristic descriptors that exhibit moderate to low discriminating power. In all but one use cases, a commercial face recognition system outperforms the characteristic descriptors. However, in low resolution surveillance camera images, some (combination of) characteristic descriptors yield better results than commercial systems.

Original language	English
Title of host publication	2016 International Conference of the Biometrics Special Interest Group (BIOSIG)
Pages	16429109
Number of pages	7
DOIs	https://doi.org/10.1109/BIOSIG.2016.7736919
Publication status	Published - 21 Sept 2016
Event	15th International Conference of the Biometrics Special Interest Group, BIOSIG 2016 - Darmstadt, Germany, Darmstadt, Germany Duration: 21 Sept 2016 → 23 Sept 2016 Conference number: 15

Conference

Conference	15th International Conference of the Biometrics Special Interest Group, BIOSIG 2016
Abbreviated title	BIOSIG 2016
Country/Territory	Germany
City	Darmstadt
Period	21/09/16 → 23/09/16
Other	21-23 Sep 2016

Keywords

2025 OA procedure
forensic use
face recognition system
evidence evaluation
FISWG characteristic descriptor
SCS-Safety
Facial feature
low resolution surveillance camera image
biometric system discriminating power

Access to Document

10.1109/BIOSIG.2016.7736919

Zeinstra16_Discriminating_PowerFinal published version, 792 KBLicence: Taverne

Cite this

@inproceedings{30a6af02075044a39d9d04ea39c60f8c,

title = "Discriminating Power of FISWG Characteristic Descriptors Under Different Forensic Use Cases",

abstract = "FISWG characteristic descriptors are facial features that can be used for evidence evaluation during forensic case work. In this paper we investigate the discriminating power of a biometric system that uses these characteristic descriptors as features under different forensic use cases. We show that in every forensic use case we can find characteristic descriptors that exhibit moderate to low discriminating power. In all but one use cases, a commercial face recognition system outperforms the characteristic descriptors. However, in low resolution surveillance camera images, some (combination of) characteristic descriptors yield better results than commercial systems.",

keywords = "2025 OA procedure, forensic use, face recognition system, evidence evaluation, FISWG characteristic descriptor, SCS-Safety, Facial feature, low resolution surveillance camera image, biometric system discriminating power",

author = "Zeinstra, \{Christopher Gerard\} and Veldhuis, \{Raymond N.J.\} and Spreeuwers, \{Lieuwe Jan\}",

year = "2016",

month = sep,

day = "21",

doi = "10.1109/BIOSIG.2016.7736919",

language = "English",

pages = "16429109",

booktitle = "2016 International Conference of the Biometrics Special Interest Group (BIOSIG)",

note = "15th International Conference of the Biometrics Special Interest Group, BIOSIG 2016, BIOSIG 2016 ; Conference date: 21-09-2016 Through 23-09-2016",

}

Zeinstra, CG, Veldhuis, RNJ & Spreeuwers, LJ 2016, Discriminating Power of FISWG Characteristic Descriptors Under Different Forensic Use Cases. in 2016 International Conference of the Biometrics Special Interest Group (BIOSIG). pp. 16429109, 15th International Conference of the Biometrics Special Interest Group, BIOSIG 2016, Darmstadt, Germany, 21/09/16. https://doi.org/10.1109/BIOSIG.2016.7736919

Discriminating Power of FISWG Characteristic Descriptors Under Different Forensic Use Cases. / Zeinstra, Christopher Gerard; Veldhuis, Raymond N.J.; Spreeuwers, Lieuwe Jan.
2016 International Conference of the Biometrics Special Interest Group (BIOSIG). 2016. p. 16429109.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Discriminating Power of FISWG Characteristic Descriptors Under Different Forensic Use Cases

AU - Zeinstra, Christopher Gerard

AU - Veldhuis, Raymond N.J.

AU - Spreeuwers, Lieuwe Jan

N1 - Conference code: 15

PY - 2016/9/21

Y1 - 2016/9/21

N2 - FISWG characteristic descriptors are facial features that can be used for evidence evaluation during forensic case work. In this paper we investigate the discriminating power of a biometric system that uses these characteristic descriptors as features under different forensic use cases. We show that in every forensic use case we can find characteristic descriptors that exhibit moderate to low discriminating power. In all but one use cases, a commercial face recognition system outperforms the characteristic descriptors. However, in low resolution surveillance camera images, some (combination of) characteristic descriptors yield better results than commercial systems.

AB - FISWG characteristic descriptors are facial features that can be used for evidence evaluation during forensic case work. In this paper we investigate the discriminating power of a biometric system that uses these characteristic descriptors as features under different forensic use cases. We show that in every forensic use case we can find characteristic descriptors that exhibit moderate to low discriminating power. In all but one use cases, a commercial face recognition system outperforms the characteristic descriptors. However, in low resolution surveillance camera images, some (combination of) characteristic descriptors yield better results than commercial systems.

KW - 2025 OA procedure

KW - forensic use

KW - face recognition system

KW - evidence evaluation

KW - FISWG characteristic descriptor

KW - SCS-Safety

KW - Facial feature

KW - low resolution surveillance camera image

KW - biometric system discriminating power

U2 - 10.1109/BIOSIG.2016.7736919

DO - 10.1109/BIOSIG.2016.7736919

M3 - Conference contribution

SP - 16429109

BT - 2016 International Conference of the Biometrics Special Interest Group (BIOSIG)

T2 - 15th International Conference of the Biometrics Special Interest Group, BIOSIG 2016

Y2 - 21 September 2016 through 23 September 2016

ER -

Discriminating Power of FISWG Characteristic Descriptors Under Different Forensic Use Cases

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this