Mask-invariant Face Recognition through Template-level Knowledge Distillation
This addresses the challenge of degraded face recognition performance due to mask-wearing, which is important for biometric security and contactless systems, though it is an incremental improvement over existing methods.
The paper tackles the problem of face recognition with masked faces, which became critical during the COVID-19 pandemic, by proposing MaskInv, a solution that uses template-level knowledge distillation to produce similar embeddings for masked and non-masked faces of the same identities. It outperforms previous state-of-the-art methods in masked vs masked and masked vs non-masked scenarios on benchmarks like MFRC-21 and MFR2, with only a minor performance loss on unmasked faces.
The emergence of the global COVID-19 pandemic poses new challenges for biometrics. Not only are contactless biometric identification options becoming more important, but face recognition has also recently been confronted with the frequent wearing of masks. These masks affect the performance of previous face recognition systems, as they hide important identity information. In this paper, we propose a mask-invariant face recognition solution (MaskInv) that utilizes template-level knowledge distillation within a training paradigm that aims at producing embeddings of masked faces that are similar to those of non-masked faces of the same identities. In addition to the distilled knowledge, the student network benefits from additional guidance by margin-based identity classification loss, ElasticFace, using masked and non-masked faces. In a step-wise ablation study on two real masked face databases and five mainstream databases with synthetic masks, we prove the rationalization of our MaskInv approach. Our proposed solution outperforms previous state-of-the-art (SOTA) academic solutions in the recent MFRC-21 challenge in both scenarios, masked vs masked and masked vs non-masked, and also outperforms the previous solution on the MFR2 dataset. Furthermore, we demonstrate that the proposed model can still perform well on unmasked faces with only a minor loss in verification performance. The code, the trained models, as well as the evaluation protocol on the synthetically masked data are publicly available: https://github.com/fdbtrs/Masked-Face-Recognition-KD.