Can contact-free measurement of heartbeat signal be used in forensics?

Publikation: Forskning - peer reviewKonferenceartikel i proceeding

Abstrakt

Biometrics and soft biometrics characteristics are of great
importance in forensics applications for identifying criminals
and law enforcement. Developing new biometrics and
soft biometrics are therefore of interest of many applications,
among them forensics. Heartbeat signals have been
previously used as biometrics, but they have been measured
using contact-based sensors. This paper extracts heartbeat
signals, using a contact-free method by a simple webcam.
The extracted signals in this case are not as precise as those
that can be extracted using contact-based sensors. However,
the contact-free extracted heartbeat signals are shown in this
paper to have some potentials to be used as soft biometrics.
Promising experimental results on a public database, have
shown that utilizing these signals can improve the accuracy
of spoofing detection in a face recognition system.
Luk

Detaljer

Biometrics and soft biometrics characteristics are of great
importance in forensics applications for identifying criminals
and law enforcement. Developing new biometrics and
soft biometrics are therefore of interest of many applications,
among them forensics. Heartbeat signals have been
previously used as biometrics, but they have been measured
using contact-based sensors. This paper extracts heartbeat
signals, using a contact-free method by a simple webcam.
The extracted signals in this case are not as precise as those
that can be extracted using contact-based sensors. However,
the contact-free extracted heartbeat signals are shown in this
paper to have some potentials to be used as soft biometrics.
Promising experimental results on a public database, have
shown that utilizing these signals can improve the accuracy
of spoofing detection in a face recognition system.
OriginalsprogEngelsk
Titel23rd European Signal Processing Conference (EUSIPCO), 2015
ForlagIEEE
Publikationsdato2015
Sider769-773
ISBN (Trykt)978-0-9928626-3-3
DOI
StatusUdgivet - 2015
PublikationsartForskning
Peer reviewJa
BegivenhedEuropean Signal Processing Conference - Nice, Frankrig
Varighed: 31 aug. 20154 sep. 2015

Konference

KonferenceEuropean Signal Processing Conference
LandFrankrig
ByNice
Periode31/08/201504/09/2015
SerieProceedings of the European Signal Processing Conference (EUSIPCO)
ISSN2076-1465

Download-statistik

Ingen data tilgængelig
ID: 212748709