Concealed Information Recognition with the Fusion of Physiological Communication Network of Facial Areas and Psychological Analysis

The interview analyst’s need to detect deception is a topic that has provided the conditions for providing solutions to empower them. So that, the experts and interview analysts can be assisted by automatically monitoring the subject's unsalient, unknown, or counterintuitive activities during the interview. The aim of this study was to combine quantitative and qualitative information to help improve the detection of deception. For this purpose, in addition to using the capacity of verbal and non-verbal analysis methods, thermal imaging technology and new methods of spatiotemporal analysis of the thermal patterns have been used to detect concealed information in individuals. Then, based on the study design, the database consisting of 48 truth-tellers and liars who participated in a mock scenario was collected. Then, two qualitative methods of verbal and non-verbal information analysis, including standard criteria-based content analysis (CBCA) and behavioral analysis interview (BAI) scoring, were used to identify liars and truth-tellers. In order to complete the obtained results based on these two methods, using effective connectivity analysis method, physiological network analysis of communication between different areas of the face was performed in thermal images of individuals. As a result of combining quantitative and qualitative information, the final accuracy of individuals' diagnosis increased from an average of 73.61% to 79.17%. The investigation of the agreement analysis between methods by kappa coefficient and analysis of confusion matrix information indicated the existence of complementary information in various quantitative and qualitative methods to identify concealed information in individuals.