Do Liars Blink Differently? Automated Blink Detection during Deceptive Interviews

Aaron C. Elkins, Nikolaos Sorros, Stefanos Zafeiriou, Judee K. Burgoon, Maja Pantic, Jay F. Nunamaker

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

250 Downloads (Pure)

Abstract

This research investigates the development and use of an automated blink detection algorithm applied to a large (N=176) deceptive interview video corpus. The automated blink detection algorithm was 93% accurate. This work represents the first analysis of deceptive blinks of this magnitude (46 hours of video) and degree of ecological validity. After applying the algorithm to the interview video corpus, deceivers were found to blink less when lying to cognitively demanding questions. In addition to deception, people blinked more over time, more when older, and less when more skilled in social expressivity. The results of this study suggest that any deception detection algorithm that relies on blinks needs to account for time, interviewee demographics and social skill, question type, and turn taking phase (interviewee listening r speaking).
Original languageEnglish
Title of host publicationProceedings of the Rapid Screening Technologies, Deception Detection and Credibility Assessment Symposium
Subtitle of host publicationJanuary 2014
EditorsMatthew Jensen, Thomas Meservy, Judee Burgoon, Jay Nunamaker
Number of pages7
Publication statusPublished - 6 Jan 2014
Externally publishedYes
EventHICSS-47 Rapid Screening Technologies, Deception Detection Credibility Assessment Symposium 2014 - Hilton Waikoloa, Big Island, United States
Duration: 6 Jan 20147 Jan 2014

Conference

ConferenceHICSS-47 Rapid Screening Technologies, Deception Detection Credibility Assessment Symposium 2014
CountryUnited States
CityBig Island
Period6/01/147/01/14

Fingerprint Dive into the research topics of 'Do Liars Blink Differently? Automated Blink Detection during Deceptive Interviews'. Together they form a unique fingerprint.

Cite this