Intelligent Deception Detection through Machine Based Interviewing

Abstract

In this paper an automatic deception detection system, which analyses participant deception risk scores from non-verbal behaviour captured during an interview conducted by an Avatar, is demonstrated. The system is built on a configuration of artificial neural networks, which are used to detect facial objects and extract non-verbal behaviour in the form of micro gestures over short periods of time. A set of empirical experiments was conducted based a typical airport security scenario of packing a suitcase. Data was collected through 30 participants participating in either a truthful or deceptive scenarios being interviewed by a machine based border guard Avatar. Promising results were achieved using raw unprocessed data on un-optimized classifier neural networks. These indicate that a machine based interviewing technique can elicit non-verbal interviewee behavior, which allows an automatic system to detect deception.

Automated Deception Detection of Males and Females From Non-Verbal Facial Micro-Gestures

“…Early protagonists of AI deception detection were keen to point that machines are not subject to fatigue and are free of human bias [6,7]. In reality there are serious concerns about bias in machine learning AI systems due to lack of diversity in the developers (the "white guy problem") [23] or poor representation of the general population in the developers or datasets [3]. This has been widely publicized in controversial arguments about the COMPASS prison release system [25].…”

Pupil Localisation and Eye Centre Estimation Using Machine Learning and Computer Vision

“…Various studies have addressed the detection and tracking of facial landmarks including the iris and pupil which has various applications-particularly, eye gaze estimation for human-machine interfaces. The control of assistive devices for disability [2], driver safety improvements [3,4], the design of diagnostic tools for brain diseases [5], cognitive research [6], automated deception detection system (ADDS) [7], and academic performance analysis [8] are some examples of such applications.…”

Cognitive Identity Management: Risks, Trust and Decisions using Heterogeneous Sources

“…The future generation security checkpoint for mass-transit hubs and large public events is anticipated to be a system that combines: biometric-enabled authentication [12], [19], [31], [45] and watchlist check [32], screening strategies [37], deception feature detection [1], [41], [43], [55], as well as concealed illicit item detection [38], [53]. Such a system will be an integral part of the security infrastructure, including logistics and surveillance network with abilities of tracking individuals-of-interest and analyzing the group behavior.…”

“…Four information sources for machine reasoning at the security checkpoint: biometrics in infrared (3−12µm), audio (70-600 Hz), and visual (400 − 700µm) spectral bands, and UWB radar illumination (3)(4)(5)(6)(7)(8)(9)(10). for children; the pitch, loudness and timbre of a human voice are the main parameters used by an e-interviewer for emotions and deception detection [1], [41], [55] [62]; -Source B: Infrared domain; the human body radiates nonvisible infrared light (3 − 12µm waves) in proportion to its temperature; this band is used for assessment of both cognitive and physical state [43]; -Source C: Visual domain, 400 − 700µm, for authentication and emotional state assessment using face and face expression recognition [31], [32]; -Source D: Radar illumination, 3-10 GHz; certain concealed items can be detected using the Ultra Wide Band (UWB) radar [24], [38].…”

“…2) There is a deep causal relationship between the above sources of information. For example, gait and behavior/emotion pattern are correlated with possible possession of illegal concealed items [1], [41], [55].…”