Zohra Rezgui - ESR3
Background:
Zohra Rezgui graduated with an MSc. in Engineering equivalent (Engineering diploma) in Statistics from the University of Carthage. Prior to that, she took intensive scientific courses program known as Higher School Preparatory Classes with a specialization in mathematics and physics at the University of Tunis in order to get access to her engineering curriculum through national exams. For her Master's thesis, she worked on facial gender recognition in unconstrained settings with the aim of describing gender representation in French TV streams at the French national institute of audiovisual. She also had an initial encounter with challenges related to privacy when she implemented a rule-based anonymization algorithm for clinical trial data during an internship at Datametrix, a Contract Research Organization. Her current research interests mainly revolve around the intersection of biometric pattern recognition with digital humanities.
Research Interests:
Biometric Pattern Recognition, Deep Learning, Digital Humanities
ESR3: Biometric profiling
Objectives:
The specific research topic of this position is to analyse and mitigate the risk of biometric profiling based on features extracted in deep-learning-based face recognition. The researcher will investigate: (1) The risk of biometric profiling by attributes that can be inferred from features, such as profession, age, gender, health condition. (2) The characteristics of the facial image that contribute to profiling. (3) Methods to reduce the risk of profiling that can be integrated in the face recognition system.
Main supervisor
Raymond Veldhuis
Co-supervisor
Farzin Deravi