Journal
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
Volume 44, Issue 10, Pages 5962-5979Publisher
IEEE COMPUTER SOC
DOI: 10.1109/TPAMI.2021.3087709
Keywords
Large-scale face recognition; additive angular margin; noisy labels; sub-class; model inversion
Funding
- University of Nottingham
- EPSRC Fellowship DEFORM: Large Scale Shape Analysis of Deformable Models of Humans [EP/S010203/1]
- FACER2VM: Face Matching for Automatic Identity Retrieval, Recognition, Verification and Management [EP/N007743/1]
- Google Faculty Award
- Imperial President's PhD Scholarship
Ask authors/readers for more resources
In this paper, the authors propose an Additive Angular Margin Loss (ArcFace) to enhance the discriminative power in face recognition. They also introduce the sub-center ArcFace method to address label noise. Additionally, they explore the inverse problem of mapping feature vectors to face images.
Recently, a popular line of research in face recognition is adopting margins in the well-established softmax loss function to maximize class separability. In this paper, we first introduce an Additive Angular Margin Loss (ArcFace), which not only has a clear geometric interpretation but also significantly enhances the discriminative power. Since ArcFace is susceptible to the massive label noise, we further propose sub-center ArcFace, in which each class contains K sub-centers and training samples only need to be close to any of the K positive sub-centers. Sub-center ArcFace encourages one dominant sub-class that contains the majority of clean faces and non-dominant sub-classes that include hard or noisy faces. Based on this self-propelled isolation, we boost the performance through automatically purifying raw web faces under massive real-world noise. Besides discriminative feature embedding, we also explore the inverse problem, mapping feature vectors to face images. Without training any additional generator or discriminator, the pre-trained ArcFace model can generate identity-preserved face images for both subjects inside and outside the training data only by using the network gradient and Batch Normalization (BN) priors. Extensive experiments demonstrate that ArcFace can enhance the discriminative feature embedding as well as strengthen the generative face synthesis.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available