Functional Biometrics – Using body-reflections as a novel class of biometric authentication systems

• Applicant: Professor Dr. Stefan Schneegass
• Subject Area: Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
• Term: from 2019 to 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 426052422

The number of systems requiring user authentication is increasing every year. Classical authentication approaches such as Personal Identification Numbers (PINs) and passwords overwhelm users since they have to remember dozens of them. Research also showed that the selection of user-generated passwords is predictable. To counteract these weaknesses, devices such as smartphones and laptops are more often equipped with biometric authentication. The most common biometric authentication approaches are fingerprints or face recognition, which use the body of a user as a physical token for identification. While these approaches currently provide a sufficient level of security, they have two inherent drawbacks: (1) the user is not able to change biometric passwords and (2) the user is leaving them essentially everywhere (e.g., leaving fingerprints by touching the environment, video surveillance cameras recording the user´s head and body). To tackle these challenges, we introduce a novel class of biometrics, called Functional Biometrics. In contrast to physical token and behavior based biometrics, Functional Biometrics exploit the user´s body as a function. Thus, this novel class of biometrics does not only rely on the body itself but also on a specific input signal generated by the system the user wants to authenticate to. This signal (e.g., audio, electrical, haptic stimulus) is modified by the user´s body, which in return generates a characteristic response through the user’s unique body reflection (e.g., propagated audio response, muscle reaction). This characteristic response operates as a biometric password. The system measures the response and compares it to a pre-stored response (i.e., the password) to authenticate the user. Functional Biometrics combine the advantages of common knowledge-based authentication approaches such as alphanumeric passwords (i.e., changeable, a multitude of passwords per user) with the advantages of biometrics authentication approaches such as fingerprints (i.e., no cognitive load and no memorability required). In this project, we will systematically explore the design space of functional biometrics. We will investigate the feasibility of different sensing and actuation technologies, develop and adapt models and algorithms for automatically authenticate users based on measured body reflection, and will develop research probes and an authentication framework to synthesize the gained knowledge.

• DFG Programme: Research Grants