Adaptive Biometric Systems in the Deep Learning Era

Giulia Orrù will present about "Adaptive Biometric Systems in the Deep Learning Era". In the previous decade, several "adaptive" systems have been proposed to address the issue of intra-class variations in face recognition (FR) systems. In particular, novel template-based self-update algorithms, capable of retaining the expressive power of a restricted number of templates across time, allow overcoming the problem of manual updating and high computational complexity of traditional self-update. Giulia's talk aims to investigate if and to what degree "optimal" self-update algorithms enhance FR performance, particularly in application scenarios where the facial trait varies dramatically over time and the high representativeness of the deep features may not be sufficient.

Curriculum vitæ

Giulia Orrù received her M.S. degree and PhD degree in Telecommunication and Computer Science Engineering from the University of Cagliari, Italy, in 2017 and 2021, respectively. She is currently an Assistant Professor of Computer Engineering at the University of Cagliari, Italy. Since 2014 she collaborates with the Pattern Recognition and Applications Laboratory research group working on pattern recognition and its applications, specifically on biometric recognition, presentation attack detection systems and adaptive biometric systems.