Published Paper Details:

Adaptive neuro-fuzzy inference system (ANFIS) training methods have gained prominence for generating robust facial key templates in privacy-preserving applications, particularly those leveraging homomorphic encryption. This review critically examines recent advancements in ANFIS training strategies, comparing their effectiveness in extracting and modelling facial features crucial for secure template generation. Emphasis is placed on the integration of ANFIS with advanced pre-processing and feature extraction techniques to enhance template accuracy and resilience against attacks. The synergy between ANFIS-based models and homomorphic encryption frameworks is explored, highlighting their potential to balance recognition performance with stringent privacy requirements. The review concludes by identifying current challenges and recommending future research directions for optimizing ANFIS training in encrypted facial recognition systems