| Description |
1. Introduction
The ability to accurately read and understand emotions and facial expressions affords one the opportunity to act
and react in an appropriate manner in any situation. This
has implications in every aspect of human interaction, including military operations and law enforcement. This valuable skill can be taught and individuals can improve their ability to read faces as well as other non-verbal communication channels (Ekman, OSullivan & Frank, 1999). This skill set is essential for situations in which there is a language barrier for example, or in situations when action is required in real-time. Prior research has concluded that only a small subset of the population is any good at distinguishing liars from truthtellers and that most perceive their detection ability to be greater than it actually is (Ekman & OSullivan, 1991). Developing training simulations and computer-based training (CBT) tools for military personnel that incorporates accurate human behavior representation (HBR) in the form of culturally relevant facial expressions and gestures specifically capturing malicious intent and deceptive clues, will greatly improve interactive outcomes.
1.1 Project Hostile Intent Introduction
The Department of Homeland Securitys Project Hostile Intent (PHI) aims to develop a non-invasive, remote, culturally independent, automated intent and deception detection system. The goals are to further validate cross-culturally, Ekmans emotional leakage theory (1993) and subsequently develop CBT and simulation tools for military, law enforcement, and security personnel. To accomplish this, a high stakes deception paradigm has been developed to determine the cues associated with lying. The Facial Action Coding System (FACS) developed by Ekman, Friesen, and Hager (2002) is being used to measure facial expressions in order to understand what corresponding emotions are displayed during deception. FACS provides a method for quantifying facial muscle movements and has been studied extensively in identifying what muscle movements correspond to the basic human emotions.
1.2 Emotional Leakage Theory
Ekmans Emotional Leakage Theory postulates that when someone is deceiving another or trying to hide some bit of information, true emotions will leak through when trying to suppress their real feelings (Frank & Ekman 1997). In other words, expressions are autonomic and micro expressions occur even without knowledge. In a deceptive scenario such as the one used in PHI, high-resolution video cameras are used to capture these micro expressions that can occur on the face for less than a quarter of a second. Preliminary findings suggest that negative emotions in the form of certain facial expressions flash on the persons face when lying or intending to deceive. These emotions are in direct contrast to what the person is conveying through speech. The key point here is that there is a discrepancy in what happens in the face that one cannot control as compared to what someone is saying or how that person should be acting in a given situation.
The theory further states that these emotions are culturally independent, meaning that even in the most remote civilizations with little to no media access, these emotions are recognized and understood (Ekman, 1999; 2004). The next phase of PHI is to replicate the current study cross-culturally. By doing this, we hope to further validate this theory and have the ability to develop very accurate automatic detectors of facial expressions and emotions that cut across cultures as well as appearances. Documenting and collecting the various gestures that are culturally specific will be an important aspect in determining behavioral cues of deceit.
2. Benefits of Using Animation for Deception Detection Training
Designing training tools that incorporate these principles of facial expression detection is critical for military personnel. An efficient method for designing these tools and simulations is to incorporate animation and create human entities that can precisely showcase the relevant muscle movements indicative of deception. Using animation will eliminate practice effects by generating different faces with various characteristics. In addition, this will support rapid tuning to reflect characteristics of an operational population. CBT has the added benefit of continued practice, which reduces decay rates and increases the learning curve. PHI will incorporate a training study evaluating the efficacy of CBT training using animation to replicate the human face and body with the FACS as the backbone.
Utilizing the FACS methodology will generate specific and accurate muscle movements that closely resemble the human face. Being able to spontaneously show the various facial expressions is a difficult task for anyone and simply posing the movements will not allow for adequate training in the real world. Creating human entities in a training simulation based on many examples of actual humans in the high-stakes deception scenario will give the trainee sufficient material to learn how to better spot inconsistencies in facial expressions and emotions related to the context of a situation. Using animation to capture micro-expressions allows for different examples with various facial structures and appearances easily and quickly to give a trainee the practice needed to master these skills.
3. References
Ekman, P. (1993). Facial expression and emotion. American Psychologist, vol 48 (4), 384-392.
Ekman, P. (1999). Facial expressions. In Dalgleish, T., & Power, M. (1999). Handbook of Cognition and Emotion. New York: John Wiley & Sons Ltd.
Ekman, P. (2004). Emotions revealed. New York: Times Books (US). London: Weidenfeld & Nicolson (world).
Ekman, P., Friesen, W.V., & Hager, J.C. (2002). The facial action coding system. Second edition. Salt Lake City: Research Nexus eBook. London: Weidenfeld & Nicolson (world).
Ekman, P., OSullivan, M. (1991). Who can catch a liar? American Psychologist, 46 (9), 913-920.
Ekman, P., OSullivan, M., & Frank, M. (1999). A few can catch a liar. Psychological Science, 10 (3), 1-6.
Frank, M. & Ekman, P. (1997). The ability to detect deceit generalizes across different types of high stakes lies. Journal of Personality and Social Psychology, vol 72 (6), 1429-1439.
Author Biography
Jennifer King is an Engineering Research Psychologist in the Advanced Information Technology (AIT) Branch at the Naval Research Laboratory (NRL). She received her Masters in Industrial Organizational Psychology from East Carolina University with a focus on web-based training research. Her current research is in support and collaboration with the Department of Homeland Security S&T Human Factors Division in developing remote, non-invasive deception detection methods incorporating various sensor technologies to determine deception and intent. Project Hostile Intent (PHI) includes a heavy emphasis on facial expression recognition utilizing state of the art computer vision technology as well as human behavior research, in order to gather relevant data to further develop models of deception. She serves a variety of roles with PHI including NRL's project coordinator and the Contracting Officer's Representative(COR). Ms. King has been trained and is certified in the Facial Action Coding System (FACS) methodology of analyzing facial expressions, which is an intregal part of PHI.
Special Thanks:
Larry Willis, PHI Project Lead: DHS S&T Human Factors Division; Dell Lunceford, Total Immersion; Dr. Mark Frank, Principle Investigator for the PHI collection site in SUNY Buffalo and deception expert; Maggie Paizan, PHI FACS Coordinator and expert FACS coder; ReallaeR; Point of Performance, LLC; The American Institutes for Research for their PHI IV&V efforts; and http://www.image-metrics.com/.
|
| Modified |
02/22/2007 00:00 |
|
|
