Advanced Techniques in Age-Invariant Face Recognition: A Survey

Abstract

Age-Invariant Face Recognition (AIFR) addresses the critical challenge of identifying individuals despite age induced facial changes over time. This paper presents a structured survey of recent advancements in AIFR, emphasizing deep learning-based solutions, particularly Generative Adversarial Networks (GANs). Through a comparative analysis of benchmark datasets MORPH-II, FG-NET, and CACD we highlight the performance trends and limitations of current models. GANs have shown exceptional promise in synthesizing realistic age- progressed or age regressed facial images, thereby improving recognition accuracy across age gaps. Further- more, we analyze key factors affecting model performance, including dataset diversity, illumination, and pose variations. Despite significant progress, challenges remain in generalizing across ethnicities, age ranges, and real-world conditions.