Biometric Intelligence in Authentication Machines: from Talking Faces to Talking Robots

Abstract

Biometric-based intelligence in authentication machines (A-machines) concerns with the problem of balancing human and machine functions in applications such as mass public events and border crossing. The focus of this paper is intelligent human-machine interfacing of behavior biometrics, which seems to play an important role in future generations of A-machines. Various pilot projects use the "talking face" technology. We suggest further improvement to A-machines based on "talking robot" technology: while interviewing the traveler, a humanoid robot recognizes faces and expressions, verifies the e-passports and the traveler (using his/her biometrics), connects with a watchlist (database), generates cognitive questions and analyzes answers, as well as detects soft biometric features in visual and infrared spectrum bands, such as temperature and blood pressure. Various parts of such technology have been implemented, and the feasibility of assembling such a multifunctional system in the form of a robot is to be studied and understood. This paper presents a survey of the related work, and also provides the authors' vision of future generations of A-machines and their applications.

Biometric-Enabled Authentication Machines: A Survey of Open-Set Real-World Applications

“…4(c)] in human authentication systems. Throughout the interview with a human, the robot processes the person's biometric data in visual and infrared bands, communicates with databases as needed, recognizes the intentions of appearance alternation or disguise (plastic surgery), level of drug and alcohol intoxication, temperature, and blood pressure (e-medicine, epidemics catastrophe), and make decisions on the necessity of manual control [46], [73], [79], [81], [82]. More advanced systems should include interview and logical reasoning, similar to those of humans.…”