Measuring cues for stand-off deception detection based on full-body nonverbal features in body-worn cameras

摘要: Deception detection is valuable in the security domain to distinguish truth from lies. It desirable many applications, such as suspect and witness interviews airport passenger screening. Interviewers are constantly trying assess credibility of a statement, usually based on intuition without objective technical support. However, psychological research has shown that humans can hardly perform better than random guessing. multi-disciplinary area with an interest different fields, psychology computer science. In last decade, several developments have helped improve accuracy lie (e.g., concealed information test, increasing cognitive load, or measurements motion capture suits) relevant cues been discovered eye blinking fiddling fingers). With presence mobile phones bodycams society, mobile, stand-off, automatic deception methodology various whole body would create new application opportunities. this paper, we study feasibility measuring these visual automatically parts body, laying groundwork for stand-off more flexible deployable sensors, body-worn cameras. We give extensive overview recent two communities: behavioral-science community special attention observed non-verbal cues, computer-vision methods able measure cues. The extracted parts: eyes, mouth, head full-body pose. performed experiment using state-of-the-art video-content-analysis (VCA) techniques quality robustly