Dr. Laurent El-Shafey proposed a general and scalable probabilistic linear discriminant analysis (PLDA) method that he applied to face and speaker recognition for
biometric applications. He investigated more particularly: (1) an exact and scalable formulation of probabilistic linear discriminant analysis (PLDA), (2) an unified framework for three session variability
modeling techniques: Inter-Session Variability modeling (ISV), Joint Factor Analysis (JFA) and Total Variability modeling (TV), (3) an efficient implementation of the proposed unified
framework, (4) the validation of the proposed framework on three tasks: face recognition, speaker recognition and gender recognition. Additionally, he developed and proposed to the
research community a new open source software library for signal processing and machine learning, as well as a set of reproducible research papers.
To download his thesis, click on the following link: Scalable Probabilistic Models for Face and Speaker Recognition