Does “lie to me” lie to you? An evaluation of facial clues to high-stakes deception

Abstract

During a forensic interview, high-stakes deception is very prevalent notwithstanding the heavy consequences that may result. Studies have shown that most untrained people cannot perform well in discerning liars and truth-tellers.Thus it has become common to adopt various technical aids to compensate for this poor judgment. Examples are polygraphs, functional Magnetic Resonance Imaging (fMRI) and linguistic analysis. However, the deception indicators used in these cases are not reliable.In the popular TV program Lie to Me, micro-expressions have been used for detecting deceit during the investigation of some criminal cases. A microexpression is considered to be a rapid and involuntary facial expression which could reveal the concealed emotion. Additionally, some psychological studies have stated that certain facial actions are more difficult to inhibit if the associated facial expressions are genuine. Similarly, these facial expressions are equally difficult to fake. This has cast light on the possibility that deception could be detected by analyzing these facial actions. However, to the best knowledge of the author, there is no computer vision research that has attempted to discriminate high-stakes deception from truth using facial expressions. Therefore, this thesis aims to test the validity of facial clues to deception detection in high-stakes situations using computer vision approaches.We note that only a limited number of the existing databases have been collected specifically for deception detection studies and none of them were obtained from real-world situations. In this thesis we present a video database of actual highstakes situations, which we have created using YouTube.We have adopted 2D appearance-based methods as the methodology to characterize the 3D facial features. Instead of building a 3D head model as is the current trend, we have extracted invariant 2D features that are related to the 3D characteristic.iii In order to discern deception and honesty, we have identified the following deceptive cues from nine separate facial regions through dynamic facial analysis: eye blink, eyebrow motion, wrinkle occurrence and mouth motion. Then these cues were integrated to form a facial behavior pattern vector. A Random Forest was trained using the collected database and applied to classify the facial patterns into deceptive and truthful categories.Despite the many uncontrolled factors (illumination, head pose and facial occlusion) in the videos in our database, we have achieved an accuracy of 76.92% when discriminating liars from truth-tellers using both micro-expressions and

Neural alignment during face‐to‐face spontaneous deception: Does gender make a difference?

“…This simultaneous processing of the same visual cues which arose from mutual eye contact of the sender and the receiver during deception, led to the increased INS in rSFC in our study (Hirsch et al, 2017; Koike et al, 2016). Moreover, previous studies showed that more eye contact not only gave liars more opportunities to deceive (Granhag & Strömwall, 2002; Jundi et al, 2013), but also made it easier for the detectors to lie detection (Su & Levine, 2016). Our study revealed the higher success rate of deception with more eye contact during the dynamic deceptive interaction, supporting the former view that eye contact is more useful for the deceiver in our task.…”

“…Frank and Ekman (2004) reported that observers' judgments depended more on facial behaviors, especially eye contact (Frank & Ekman, 2004). Maintaining eye contact could improve observers' ability to discern the truth, which makes lie‐detection easier and more accurate (Su & Levine, 2016; Vrij et al, 2010). Although many previous studies have so far revealed the key role of eye contact in deception, little is known about the details of gender difference in eye contact during deception.…”

Airport artificial intelligence can detect deception: or am i lying?

“…Popular media such as Lie to Me has seen an increase in characters like Dr Cal Lightman and his 'wizard' type abilities to detect deception (Levine, Serota, & Shulman, 2010;Su & Levine, 2016) and the portrayal of advanced technological systems as seen in CSI (Baskin & Sommers, 2010) has led many to overestimate what science can and cannot do.…”

Do Lie Detection Tools Really Catch Liars?

“…Matsumoto and Hwang concluded that the weak findings of past studies can be explained by the lack of detailed assessment of (micro)expressions at different durations, as they did in their study 81 . Furthermore, Su and Levine provide evidence that, through a machine learning analyzing a combination of micro and macroexpressions (normal duration expressions), they achieved a 76.92% accuracy in distinguishing liars and truthtellers 82 . However, there was a lack of ground truth in Su and Levine's study, and their machine learning approach showed that macroexpressions have much more predictive value than micro-expressions 82 .…”