Topic: Pose Impact Estimation on Face Recognition using 3D-Aware Synthetic Data with Application to Quality Assessment
by Marcel Grimmer
D19/2.03a, February 23, 2023 (Thursday), 12.00 noon
Keywords — Face recogniton, Synthetic Data, Quality Assessment
Across the globe, face recognition (FR) systems have become increasingly prevalent in our everyday life. In the European Union, sensitive applications are equipped with biometric recognition, such as the Entry-Exit System (EES). Evaluating the quality of facial images is essential for operating face recognition systems with sufficient accuracy. The recent advances in face quality standardisation ISO/IEC WD 29794-5 recommend the usage of component quality measures for breaking down face quality into its individual factors, hence providing valuable feedback for operators to re-capture low-quality images.
In light of recent advances in 3D-aware generative adversarial networks, we propose a novel dataset, „Syn-YawPitch“, comprising 1,000 identities with varying yaw-pitch angle combinations. Utilizing this dataset, we demonstrate that pitch angles beyond 30 degrees have a significant impact on the biometric performance of current face recognition systems. With this knowledge, we propose a lightweight pose quality estimator that adheres to the standards of ISO/IEC WD 29794-5.