Pain Expression as a Biometric: Why Patients’ Self-Reported Pain doesn’t Match with the Objectively Measured Pain?

Publikation: Forskning - peer reviewKonferenceartikel i proceeding

Abstrakt

Developing a vision-based efficient and automatic pain intensity measurement system requires the understanding of the relationship between self-reported pain intensity and pain expression in the facial videos. In this paper, we first demonstrate how pain expression in facial video frames may not match with the self-reported score. This is because the pain and non-pain frames are not always visually distinctive; though the self-report tells different story of having pain and non-pain status. On the other hand previous studies reported that general facial expressions can be used as biometrics. Thus, in this paper we investigated the relevance of pain expression from facial video to be used as a biometric or soft-biometric trait. In order to do that, we employed a biometric person recognition scenario by using features obtained from the pain expression pattern found in the temporal axis of subjects’ videos. The results confirmed that the pain expression patterns have distinctive features between the subjects of the UNBC McMaster shoulder pain database. We concluded that as the pain expression patterns have subjective features as a biometric, this can also cause the difference between self-reported pain level and the visually observed pain intensity level.
Luk

Detaljer

Developing a vision-based efficient and automatic pain intensity measurement system requires the understanding of the relationship between self-reported pain intensity and pain expression in the facial videos. In this paper, we first demonstrate how pain expression in facial video frames may not match with the self-reported score. This is because the pain and non-pain frames are not always visually distinctive; though the self-report tells different story of having pain and non-pain status. On the other hand previous studies reported that general facial expressions can be used as biometrics. Thus, in this paper we investigated the relevance of pain expression from facial video to be used as a biometric or soft-biometric trait. In order to do that, we employed a biometric person recognition scenario by using features obtained from the pain expression pattern found in the temporal axis of subjects’ videos. The results confirmed that the pain expression patterns have distinctive features between the subjects of the UNBC McMaster shoulder pain database. We concluded that as the pain expression patterns have subjective features as a biometric, this can also cause the difference between self-reported pain level and the visually observed pain intensity level.
OriginalsprogEngelsk
TitelIEEE International Conference on Identity, Security and Behavior Analysis 2017
UdgiverIEEE Computer Society Press
Publikationsdato2017
StatusAccepteret/In press - 2017
Begivenhed - New Delhi, Indien

Konference

KonferenceIEEE International Conference on Identity, Security and Behavior Analysis
LandIndien
ByNew Delhi
Periode22/02/201724/02/2017

Download-statistik

Ingen data tilgængelig
ID: 245553204