Heartbeat Signal from Facial Video for Biometric Recognition

Publication: Research - peer-reviewArticle in proceeding

Abstract

Different biometric traits such as face appearance and heartbeat signal from Electrocardiogram (ECG)/Phonocardiogram (PCG) are widely used in the human identity recognition. Recent advances in facial video based measurement of cardio-physiological parameters such as heartbeat rate, respiratory rate, and blood volume pressure provide the possibility of extracting heartbeat signal from facial video instead of using obtrusive ECG or PCG sensors in the body. This paper proposes the Heartbeat Signal from Facial Video (HSFV) as a new biometric trait for human identity recognition, for the first time to the best of our knowledge. Feature extraction from the HSFV is accomplished by employing Radon transform on a waterfall model of the replicated HSFV. The pairwise Minkowski distances are obtained from the Radon image as the features. The authentication is accomplished by a decision tree based supervised approach. The potential of the proposed HSFV biometric for human identification is demonstrated on a public database.
Close

Details

Different biometric traits such as face appearance and heartbeat signal from Electrocardiogram (ECG)/Phonocardiogram (PCG) are widely used in the human identity recognition. Recent advances in facial video based measurement of cardio-physiological parameters such as heartbeat rate, respiratory rate, and blood volume pressure provide the possibility of extracting heartbeat signal from facial video instead of using obtrusive ECG or PCG sensors in the body. This paper proposes the Heartbeat Signal from Facial Video (HSFV) as a new biometric trait for human identity recognition, for the first time to the best of our knowledge. Feature extraction from the HSFV is accomplished by employing Radon transform on a waterfall model of the replicated HSFV. The pairwise Minkowski distances are obtained from the Radon image as the features. The authentication is accomplished by a decision tree based supervised approach. The potential of the proposed HSFV biometric for human identification is demonstrated on a public database.
Original languageEnglish
Title of host publicationImage Analysis : 19th Scandinavian Conference, SCIA 2015, Copenhagen, Denmark, June 15-17, 2015. Proceedings
Number of pages12
PublisherSpringer
Publication date2015
Pages165-174
ISBN (print)978-3-319-19664-0
ISBN (electronic)978-3-319-19665-7
DOI
StatePublished - 2015
EventScandinavian Conference on Image Analysis - Copenhagen, Denmark

Conference

ConferenceScandinavian Conference on Image Analysis
Nummer19
LandDenmark
ByCopenhagen
Periode15/06/201517/06/2015
SeriesLecture Notes in Computer Science
Volume9127
ISSN0302-9743

    Keywords

  • biometric, identification, Radon transform, heartbeat, facial video

Download statistics

No data available
ID: 208955506