ABSTRACT
The globalization pandemic of COVID-19 has made wearing masks to become a norm in people's lives, and this preventive measure brings new challenges to face recognition algorithms. To address this problem, in this paper, a multi-branch network is proposed to simultaneously complete the task of the masked face detection and recognition. Firstly, this network improves the Swin Transformer for extraction of facial features. Secondly, a face organ attention mechanism, FOA, is proposed to make the model focus on the face organs that are not covered by masks. Then, in order to overcome the problem of inadequate masked face dataset, a data augmentation method using 3D face mesh is proposed to add face mask. Finally, the experimental results show that, compared with the benchmark model, the proposed model reduces the number of model parameters by 60.6%, while the AP of masked face recognition increases by 5.33%, which better balances speed and accuracy. © 2022 IEEE.
ABSTRACT
Masked face recognition, a non-contact biometric technology, has attracted much attention and developed rapidly during the coronavirus disease 2019 (COVID-19) outbreak. The existing work trains the masked face recognition model based on a large number of 2D masked face images. However, in practical application scenarios, it is difficult to obtain a large number of masked face images in a short period of time. Therefore, combined with 3D face recognition technology, this paper proposes a masked face recognition model trained with non-masked face images. In this paper, we locate and segment the complete face region and the face region not occluded by masks from the face point clouds. The geometric features of the 3D face surface, namely depth, azimuth, and elevation, are extracted from the above two regions to generate training data. The proposed masked face recognition model based on vision Transformer divides the complete faces and part of the faces into sequence images, and then captures the relationship between the image slices to compensate for the impact caused by the lack of face information, thereby improving the recognition performance. Comparative experiments with the state-of-the-art masked face recognition work are carried out on four databases. The experimental results show that the recognition accuracy of the proposed model is improved by 9.86% on Bosphorus database, 16.77% on CASIA-3D FaceV1 database, 2.32% on StirlingESRC database, and 34.81% on Ajmal main database, respectively, which verifies the effectiveness and stability of the proposed model. © 2022 IEEE.
ABSTRACT
During the coronavirus pandemic, the demand for contactless biometrics technology has promoted the development of masked face recognition. Training a masked face recognition model needs to address two crucial issues: a lack of large-scale realistic masked face datasets and the difficulty of obtaining robust face representations due to the huge difference between complete faces and masked faces. To tackle with the first issue, this paper proposes to train a 3D masked face recognition network with non-masked face images. For the second issue, this paper utilizes the geometric features of 3D face, namely depth, azimuth, and elevation, to represent the face. The inherent advantages of 3D face enhance the stability and practicability of 3D masked face recognition network. In addition, a facial geometry extractor is proposed to highlight discriminative facial geometric features so that the 3D masked face recognition network can take full advantage of the depth, azimuth and elevation information in distinguishing face identities. The experimental results on four public 3D face datasets show that the proposed 3D masked face recognition network improves the accuracy of the masked face recognition, which verifies the feasibility of training the masked face recognition model with non-masked face images. © 2022 ACM.