Controllable and Guided Face Synthesis for Unconstrained Face Recognition
This addresses the problem of unconstrained face recognition for applications like surveillance or security, but it is incremental as it builds on existing synthesis and recognition methods.
The paper tackles the domain gap between semi-constrained training and unconstrained testing in face recognition by proposing a controllable face synthesis model that mimics target dataset distributions, resulting in a 5.76% Rank1 improvement on benchmarks like IJB-B and IJB-C.
Although significant advances have been made in face recognition (FR), FR in unconstrained environments remains challenging due to the domain gap between the semi-constrained training datasets and unconstrained testing scenarios. To address this problem, we propose a controllable face synthesis model (CFSM) that can mimic the distribution of target datasets in a style latent space. CFSM learns a linear subspace with orthogonal bases in the style latent space with precise control over the diversity and degree of synthesis. Furthermore, the pre-trained synthesis model can be guided by the FR model, making the resulting images more beneficial for FR model training. Besides, target dataset distributions are characterized by the learned orthogonal bases, which can be utilized to measure the distributional similarity among face datasets. Our approach yields significant performance gains on unconstrained benchmarks, such as IJB-B, IJB-C, TinyFace and IJB-S (+5.76% Rank1).