A Video-Based Screening System for Automated Risk Assessment Using Nuanced Facial Features

• 成果类型: Article
• 署名作者: Pentland, Steven J.; Twyman, Nathan W.; Burgoon, Judee K.; Nunamaker, Jay F., Jr.; Diller, Christopher B. R.
• 署名单位: University of Arizona; University of Missouri System; Missouri University of Science & Technology; University of Arizona; University of Arizona; National Science Foundation (NSF); University of Arizona; University of Arizona; University of Nebraska System
• 刊物名称: JOURNAL OF MANAGEMENT INFORMATION SYSTEMS
• ISSN/ISSBN: 0742-1222
• DOI: 10.1080/07421222.2017.1393304
• 发表日期: 2017
• 页码: 970-993
• 关键词: Deception; INFORMATION; emotions; CLUES; lies; cues; me

摘要

This study investigates the development of an automated interviewing system that uses facial behavior as an indicator of the risk of given illicit behavior. Traditional facial emotion indicators of risk in semistructured dialogue may have limitations in an automated approach. However, an initial analysis of mock crime interviews suggests that the face may exhibit some form of rigidity during highly structured interviews. An interviewing system design using facial rigidity analysis was implemented and experimentally evaluated, the results of which further reveal that the rigidity is fairly generalized across the face. Whereas existing theory traditionally focuses on leakage of facial expressions, this study provides evidence that neutralization of facial expression may be a valuable alternative for automated interviewing systems. The proof-of-concept system in this study may help human risk assessment move beyond traditional boundaries, into fields such as auditing, emergency room management, and security screening.