# A review of calibration methods for biometric systems in forensic applications

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

41 Downloads (Pure)

### Abstract

When, in a criminal case there are traces from a crime scene - e.g., finger marks or facial recordings from a surveillance camera - as well as a suspect, the judge has to accept either the hypothesis \emph{$H_{p}$} of the prosecution, stating that the trace originates from the subject, or the hypothesis of the defense \emph{$H_d$}, stating the opposite. The current practice is that forensic experts provide a degree of support for either of the two hypotheses, based on their examinations of the trace and reference data - e.g., fingerprints or photos - taken from the suspect. There is a growing interest in a more objective quantitative support for these hypotheses based on the output of biometric systems instead of manual comparison. However, the output of a score-based biometric system is not directly suitable for quantifying the evidential value contained in a trace. A suitable measure that is gradually becoming accepted in the forensic community is the Likelihood Ratio (LR) which is the ratio of the probability of evidence given \emph{$H_p$} and the probability of evidence given \emph{$H_d$}. In this paper we study and compare different score-to-LR conversion methods (called calibration methods). We include four methods in this comparative study: Kernel Density Estimation (KDE), Logistic Regression (Log Reg), Histogram Binning (HB), and Pool Adjacent Violators (PAV). Useful statistics such as mean and bias of the bootstrap distribution of \emph{LRs} for a single score value are calculated for each method varying population sizes and score location.
Original language English 33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012 Enschede Werkgemeenschap voor Informatie- en Communicatietheorie (WIC) 126-133 8 978-90-365-3383-6 Published - May 2012 33rd WIC Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012 - Boekelo, NetherlandsDuration: 24 May 2012 → 25 May 2012Conference number: 33

### Conference

Conference 33rd WIC Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012 Netherlands Boekelo 24/05/12 → 25/05/12

Biometrics
Calibration
Crime
Logistics
Cameras
Statistics

• METIS-287889
• EWI-21962
• SCS-Safety
• IR-80693

### Cite this

Ali, T., Spreeuwers, L., & Veldhuis, R. (2012). A review of calibration methods for biometric systems in forensic applications. In 33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012 (pp. 126-133). Enschede: Werkgemeenschap voor Informatie- en Communicatietheorie (WIC).
Ali, Tauseef ; Spreeuwers, Luuk ; Veldhuis, Raymond. / A review of calibration methods for biometric systems in forensic applications. 33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012. Enschede : Werkgemeenschap voor Informatie- en Communicatietheorie (WIC), 2012. pp. 126-133
@inproceedings{696fcc43755e4ee298fd4d63dc09c188,
title = "A review of calibration methods for biometric systems in forensic applications",
abstract = "When, in a criminal case there are traces from a crime scene - e.g., finger marks or facial recordings from a surveillance camera - as well as a suspect, the judge has to accept either the hypothesis \emph{$H_{p}$} of the prosecution, stating that the trace originates from the subject, or the hypothesis of the defense \emph{$H_d$}, stating the opposite. The current practice is that forensic experts provide a degree of support for either of the two hypotheses, based on their examinations of the trace and reference data - e.g., fingerprints or photos - taken from the suspect. There is a growing interest in a more objective quantitative support for these hypotheses based on the output of biometric systems instead of manual comparison. However, the output of a score-based biometric system is not directly suitable for quantifying the evidential value contained in a trace. A suitable measure that is gradually becoming accepted in the forensic community is the Likelihood Ratio (LR) which is the ratio of the probability of evidence given \emph{$H_p$} and the probability of evidence given \emph{$H_d$}. In this paper we study and compare different score-to-LR conversion methods (called calibration methods). We include four methods in this comparative study: Kernel Density Estimation (KDE), Logistic Regression (Log Reg), Histogram Binning (HB), and Pool Adjacent Violators (PAV). Useful statistics such as mean and bias of the bootstrap distribution of \emph{LRs} for a single score value are calculated for each method varying population sizes and score location.",
keywords = "METIS-287889, EWI-21962, SCS-Safety, IR-80693",
author = "Tauseef Ali and Luuk Spreeuwers and Raymond Veldhuis",
year = "2012",
month = "5",
language = "English",
isbn = "978-90-365-3383-6",
pages = "126--133",
booktitle = "33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012",
publisher = "Werkgemeenschap voor Informatie- en Communicatietheorie (WIC)",
address = "Netherlands",

}

Ali, T, Spreeuwers, L & Veldhuis, R 2012, A review of calibration methods for biometric systems in forensic applications. in 33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012. Werkgemeenschap voor Informatie- en Communicatietheorie (WIC), Enschede, pp. 126-133, 33rd WIC Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012, Boekelo, Netherlands, 24/05/12.
33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012. Enschede : Werkgemeenschap voor Informatie- en Communicatietheorie (WIC), 2012. p. 126-133.

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

TY - GEN

T1 - A review of calibration methods for biometric systems in forensic applications

AU - Ali, Tauseef

AU - Spreeuwers, Luuk

AU - Veldhuis, Raymond

PY - 2012/5

Y1 - 2012/5

N2 - When, in a criminal case there are traces from a crime scene - e.g., finger marks or facial recordings from a surveillance camera - as well as a suspect, the judge has to accept either the hypothesis \emph{$H_{p}$} of the prosecution, stating that the trace originates from the subject, or the hypothesis of the defense \emph{$H_d$}, stating the opposite. The current practice is that forensic experts provide a degree of support for either of the two hypotheses, based on their examinations of the trace and reference data - e.g., fingerprints or photos - taken from the suspect. There is a growing interest in a more objective quantitative support for these hypotheses based on the output of biometric systems instead of manual comparison. However, the output of a score-based biometric system is not directly suitable for quantifying the evidential value contained in a trace. A suitable measure that is gradually becoming accepted in the forensic community is the Likelihood Ratio (LR) which is the ratio of the probability of evidence given \emph{$H_p$} and the probability of evidence given \emph{$H_d$}. In this paper we study and compare different score-to-LR conversion methods (called calibration methods). We include four methods in this comparative study: Kernel Density Estimation (KDE), Logistic Regression (Log Reg), Histogram Binning (HB), and Pool Adjacent Violators (PAV). Useful statistics such as mean and bias of the bootstrap distribution of \emph{LRs} for a single score value are calculated for each method varying population sizes and score location.

AB - When, in a criminal case there are traces from a crime scene - e.g., finger marks or facial recordings from a surveillance camera - as well as a suspect, the judge has to accept either the hypothesis \emph{$H_{p}$} of the prosecution, stating that the trace originates from the subject, or the hypothesis of the defense \emph{$H_d$}, stating the opposite. The current practice is that forensic experts provide a degree of support for either of the two hypotheses, based on their examinations of the trace and reference data - e.g., fingerprints or photos - taken from the suspect. There is a growing interest in a more objective quantitative support for these hypotheses based on the output of biometric systems instead of manual comparison. However, the output of a score-based biometric system is not directly suitable for quantifying the evidential value contained in a trace. A suitable measure that is gradually becoming accepted in the forensic community is the Likelihood Ratio (LR) which is the ratio of the probability of evidence given \emph{$H_p$} and the probability of evidence given \emph{$H_d$}. In this paper we study and compare different score-to-LR conversion methods (called calibration methods). We include four methods in this comparative study: Kernel Density Estimation (KDE), Logistic Regression (Log Reg), Histogram Binning (HB), and Pool Adjacent Violators (PAV). Useful statistics such as mean and bias of the bootstrap distribution of \emph{LRs} for a single score value are calculated for each method varying population sizes and score location.

KW - METIS-287889

KW - EWI-21962

KW - SCS-Safety

KW - IR-80693

M3 - Conference contribution

SN - 978-90-365-3383-6

SP - 126

EP - 133

BT - 33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012

PB - Werkgemeenschap voor Informatie- en Communicatietheorie (WIC)

CY - Enschede

ER -

Ali T, Spreeuwers L, Veldhuis R. A review of calibration methods for biometric systems in forensic applications. In 33rd Symposium on Information Theory in the Benelux and the 2nd Joint WIC/IEEE Symposium on Information Theory and Signal Processing in the Benelux 2012. Enschede: Werkgemeenschap voor Informatie- en Communicatietheorie (WIC). 2012. p. 126-133