A research team led by Professor Junsuk Rho of the Departments of Mechanical Engineering, Chemical Engineering, Electrical Engineering, and the Graduate School of Convergence Science and Technology at POSTECH, together with Dr. Seokho Lee and integrated M.S.–Ph.D. student Beomseok Oh of the Department of Mechanical Engineering, has developed a new biometric authentication technology that can identify users without storing their fingerprint information as digital data. The study was published in the international journal Nature Sensors.
Fingerprint recognition is widely used in everyday applications, from unlocking smartphones to financial transactions and access control. Unlike passwords, however, fingerprints cannot easily be changed once compromised. Conventional biometric authentication systems first convert a fingerprint into digital data and then compare it with previously stored fingerprint information. Even when the stored information is encrypted, the authentication process still requires access to biometric data, leaving potential security vulnerabilities.
Rather than seeking a safer way to store fingerprint information, the researchers approached the problem by eliminating the process of digitally storing and comparing fingerprints itself. Instead of sending fingerprint data to a computer for analysis, the system completes the required computation through the physical interaction between ultrasound and the fingerprint. To achieve this, the team proposed a metasurface-driven loop-diffractive neural network, or loop-DNN. A metasurface is an artificial structure capable of precisely controlling the direction, intensity, and phase of ultrasound and, in this study, functions similarly to elements in an artificial neural network.
The researchers designed the system so that ultrasound passing through the metasurface is directed to different locations depending on the fingerprint. When a registered fingerprint is presented, the acoustic energy is concentrated in a "registered" region, whereas an unregistered fingerprint directs the energy toward a "non-registered" region. Authentication can therefore be determined simply by comparing the signal intensities in the two regions, without reconstructing a fingerprint image or extracting fingerprint features through separate digital computation. The researchers further controlled the phase, amplitude, and switch states of the ultrasound, allowing a single metasurface to process fingerprint information with expanded degrees of freedom for wave manipulation. Whereas conventional diffractive neural networks mainly modulate phase, this work additionally controls the intensity and transmission (on/off) of the ultrasound, enabling more versatile use of the waves.
In experiments using fingerprints from four individuals, the system achieved an accuracy of 97.92% across a total of 192 validation samples. Additional numerical analysis using an international fingerprint recognition database demonstrated an accuracy of up to 99.75% for a single registered user. The researchers also observed a general improvement in recognition performance as the design degrees of freedom of the metasurface increased.
The significance of this study lies in shifting biometric authentication from a "store-and-compare" approach to one based on "physical computation and direct decision-making." Because fingerprint information does not need to be stored separately, the approach could reduce the risk of biometric data leakage. The technology may ultimately find applications in fields requiring high levels of security, including finance, healthcare, and defense, as well as in automobiles, smartphones, and wearable devices.
For practical implementation, however, the current experimental system operating at approximately 40 kHz must be miniaturized toward the MHz frequency range used in industrial sensors, and the technology must be further developed to support multiple registered users. The research team plans to advance the platform by combining multilayer metasurfaces with digital technologies.
Professor Junsuk Rho, who led the study, said, "Rather than developing a faster algorithm or stronger encryption technology, this study changes the starting point of biometric authentication itself. By eliminating the need to separately store biometric information, we expect this approach to develop into a new authentication method that can reduce the risk of fingerprint information being exposed through hacking."
This research was supported by the POSCO Holdings N.EX.T Impact Program, the Mid-Career Researcher Program of the Ministry of Science and ICT and the National Research Foundation of Korea, the Future-Pioneering Convergence Technology Program, the Doctoral Student Research Incentive Program of the Ministry of Education and the National Research Foundation of Korea, and the InnoCORE Program of the Ministry of Science and ICT.
