Shuhui Wang, Xiaochen Mao
Shanghai Institute of Measurement and Testing Technology, 1500 Zhang Heng Road, Shanghai, 201203, P.R.China.
Procedia Comput Sci. 2022;208:145-151. doi: 10.1016/j.procs.2022.10.022. Epub 2022 Nov 2.
With the recent worldwide COVID-19 pandemic, almost everyone wears a mask daily, leading to severe degradation in the accuracy of conventional face recognition systems. Several works improve the performance of masked faces by adopting synthetic masked face images for training. However, such methods often cause performance degradation on unmasked faces, raising the contradiction between the face recognition system's accuracy on unmasked and masked faces. In this paper, we propose a dual-proxy face recognition training method to improve masked faces' performance while maintaining unmasked faces' performance. Specifically, we design two fully-connected layers as the unmasked and masked feature space proxies to alleviate the significant difference between the two data distributions. The cross-space constraints are adopted to ensure the intra-class compactness and inter-class discrepancy. Extensive experiments on popular unmasked face benchmarks and masked face benchmarks, including real-world mask faces and the generated mask faces, demonstrate our method's superiority over the state-of-the-art methods on masked faces without incurring a notable accuracy degradation on unmasked faces.
随着近期全球新冠疫情的爆发，几乎每个人每天都佩戴口罩，这导致传统人脸识别系统的准确率严重下降。一些工作通过采用合成的蒙面人脸图像进行训练来提高蒙面人脸的识别性能。然而，这类方法往往会导致未蒙面人脸的识别性能下降，从而引发了人脸识别系统在未蒙面和蒙面人脸识别准确率之间的矛盾。在本文中，我们提出了一种双代理人脸识别训练方法，以在保持未蒙面人脸识别性能的同时提高蒙面人脸的识别性能。具体而言，我们设计了两个全连接层作为未蒙面和蒙面特征空间代理，以缓解两种数据分布之间的显著差异。采用跨空间约束来确保类内紧凑性和类间差异性。在流行的未蒙面人脸基准和蒙面人脸基准上进行的大量实验，包括真实世界的戴口罩人脸和生成的戴口罩人脸，证明了我们的方法在蒙面人脸上优于现有方法，同时不会导致未蒙面人脸的准确率显著下降。
