Abstract
The softmax-based loss function and its variants (e.g., cosface, sphereface, and arcface) significantly improve the face recognition performance in wild unconstrained scenes. A common practice of these algorithms is to perform optimizations on the multiplication between the embedding features and the linear transformation matrix. However in most cases, the dimension of embedding features is given based on traditional design experience, and there is less-studied on improving performance using the feature itself when giving a fixed size. To address this challenge, this paper presents a softmax approximation method called SubFace, which employs the subspace feature to promote the performance of face recognition. Specifically, we dynamically select the non-overlapping subspace features in each batch during training, and then use the subspace features to approximate full-feature among softmax-based loss, so the discriminability of the deep model can be significantly enhanced for face recognition. Comprehensive experiments conducted on benchmark datasets demonstrate that our method can significantly improve the performance of vanilla CNN baseline, which strongly proves the effectiveness of the subspace strategy with the margin-based loss, e.g. ArcFace with our strategy can achieve the best performance of 99.85% and 93.48% on LFW and CPLFW dataset respectively.
Similar content being viewed by others
Availability of data and material
The data used for the experiments in this paper are available online, see Section 4.1 for more details.
Code Availibility
The code will be publicly available once the work is published upon agreement of different sides.
Notes
The feature sampling rate of 1.0 represents the original training strategy.
References
An X, Zhu X, Gao Y et al (2021) Partial fc: Training 10 million identities on a single machine. In: IEEE International conference on computer vision, pp 1445–1449
Chen S, Liu Y, Gao X et al (2018) Mobilefacenets: Efficient cnns for accurate real-time face verification on mobile devices. In: Chinese conference on biometric recognition, Springer, pp 428–438
Chopra S, Hadsell R, LeCun Y (2005) Learning a similarity metric discriminatively, with application to face verification. In: IEEE Conference on computer vision and pattern recognition, IEEE, pp 539–546
Deng J, Zhou Y, Zafeiriou S (2017) Marginal loss for deep face recognition. In: IEEE Conference on computer vision and pattern recognition workshops, pp 60–68
Deng J, Guo J, Xue N et al (2019) Arcface: Additive angular margin loss for deep face recognition. In: IEEE Conference on computer vision and pattern recognition, pp 4690–4699
Deng J, Guo J, Liu T et al (2020a) Sub-center arcface: Boosting face recognition by large-scale noisy web faces. In: European conference on computer vision, Springer, pp 741–757
Deng J, Guo J, Ververas E et al (2020b) Retinaface: Single-shot multi-level face localisation in the wild. In: IEEE Conference on computer vision and pattern recognition, pp 5203–5212
Ding C, Tao D (2015) Robust face recognition via multimodal deep face representation. IEEE Trans Multimed 17(11):2049–2058
Guo Y, Zhang L, Hu Y et al (2016) Ms-celeb-1m: A dataset and benchmark for large-scale face recognition. In: European conference on computer vision, Springer, pp 87–102
Han D, Kim J, Kim J (2017) Deep pyramidal residual networks. In: IEEE Conference on computer vision and pattern recognition, pp 5927–5935
He K, Zhang X, Ren S et al (2016) Deep residual learning for image recognition. In: IEEE Conference on computer vision and pattern recognition, pp 770–778
Huang GB, Mattar M, Berg T et al (2008a) Labeled faces in the wild: A database forstudying face recognition in unconstrained environments. In: Workshop on faces in’Real-Life’Images: detection, alignment, and recognition
Huang GB, Mattar M, Berg T et al (2008b) Labeled faces in the wild: A database forstudying face recognition in unconstrained environments. In: Workshop on faces in’Real-Life’Images: detection, alignment, and recognition
Huang Y, Wang Y, Tai Y et al (2020) Curricularface: adaptive curriculum learning loss for deep face recognition. In: IEEE Conference on computer vision and pattern recognition, pp 5901–5910
Kemelmacher-Shlizerman I, Seitz SM, Miller D et al (2016) The megaface benchmark: 1 million faces for recognition at scale. In: IEEE Conference on computer vision and pattern recognition, pp 4873–4882
Kim Y, Park W, Roh MC et al (2020) Groupface: Learning latent groups and constructing group-based representations for face recognition. In: IEEE Conference on computer vision and pattern recognition, pp 5621–5630
Liu B, Deng W, Zhong Y et al (2019a) Fair loss: Margin-aware reinforcement learning for deep face recognition. In: IEEE Conference on Computer Vision and Pattern Recognition, pp 10,052–10,061
Liu H, Zhu X, Lei Z et al (2019b) Adaptiveface: Adaptive margin and sampling for face recognition. In: IEEE Conference on computer vision and pattern recognition, pp 11,947–11,956
Liu W, Wen Y, Yu Z et al (2016) Large-margin softmax loss for convolutional neural networks. In: International conference on machine learning, pp 7
Liu W, Wen Y, Yu Z et al (2017) Sphereface: Deep hypersphere embedding for face recognition. In: IEEE Conference on computer vision and pattern recognition, pp 212–220
Martindez-Diaz Y, Luevano LS, Mendez-Vazquez H et al (2019) Shufflefacenet: A lightweight face architecture for efficient and highly-accurate face recognition. In: IEEE International conference on computer vision workshops, pp 0–0
Martinez-Diaz Y, Nicolas-Diaz M, Mendez-Vazquez H et al (2021) Benchmarking lightweight face architectures on specific face recognition scenarios. Artif Intell Rev 54(8):6201–6244
Maze B, Adams J, Duncan JA et al (2018) Iarpa janus benchmark-c: Face dataset and protocol. In: International conference on biometrics, IEEE, pp 158–165
Moschoglou S, Papaioannou A, Sagonas C et al (2017) Agedb: the first manually collected, in-the-wild age database. In: IEEE Conference on computer vision and pattern recognition workshops, pp 51–59
Ng HW, Winkler S (2014) A data-driven approach to cleaning large face datasets. In: IEEE International conference on image processing, IEEE, pp 343–347
Paszke A, Gross S, Massa F et al (2019) Pytorch: An imperative style, high-performance deep learning library. In: Advances in neural information processing systems 32
Qian Q, Shang L, Sun B et al (2019) Softtriple loss: Deep metric learning without triplet sampling. In: IEEE International conference on computer vision, pp 6450–6458
Sankaranarayanan S, Alavi A, Castillo CD et al (2016) Triplet probabilistic embedding for face verification and clustering. IEEE International Conference on Biometrics Theory. Applications and Systems, IEEE, pp 1–8
Schroff F, Kalenichenko D, Philbin J (2015) Facenet: A unified embedding for face recognition and clustering. In: IEEE Conference on computer vision and pattern recognition, pp 815–823
Sohn K (2016) Improved deep metric learning with multi-class n-pair loss objective. In: Advances in neural information processing systems 29
Sun Y, Chen Y, Wang X et al (2014) Deep learning face representation by joint identification-verification. In: Advances in neural information processing systems 27
Sun Y, Wang X, Tang X (2016) Sparsifying neural network connections for face recognition. In: IEEE Conference on computer vision and pattern recognition, pp 4856–4864
Wang F, Cheng J, Liu W et al (2018) Additive margin softmax for face verification. IEEE Signal Process Lett 25(7):926–930
Wang H, Wang Y, Zhou Z et al (2018b) Cosface: Large margin cosine loss for deep face recognition. In: IEEE Conference on computer vision and pattern recognition, pp 5265–5274
Wen Y, Zhang K, Li Z et al (2016) A discriminative feature learning approach for deep face recognition. In: European conference on computer vision, Springer, pp 499–515
Whitelam C, Taborsky E, Blanton A et al (2017) Iarpa janus benchmark-b face dataset. In: IEEE Conference on computer vision and pattern recognition workshops, pp 90–98
Xiang S, Fu Y, Chen H et al (2020a) Multi-level feature learning with attention for person re-identification. Multimed Tools Appl 79(43):32,079–32,093
Xiang S, Fu Y, Xie M et al (2020b) Unsupervised person re-identification by hierarchical cluster and domain transfer. Multimed Tools Appl 79:19,769–19,786
Xiang S, Fu Y, You G et al (2020c) Unsupervised domain adaptation through synthesis for person re-identification. In: 2020 IEEE International conference on multimedia and expo (ICME), IEEE, pp 1–6
Xiang S, Fu Y, Guan M et al (2022) Learning from self-discrepancy via multiple co-teaching for cross-domain person re-identification. Mach Learn pp 1–18
Xiang S, Chen H, Ran W et al (2023a) Deep multimodal representation learning for generalizable person re-identification. Mach Learn pp 1–19
Xiang S, Qian D, Guan M et al (2023) Less is more: Learning from synthetic data with fine-grained attributes for person re-identification. ACM Trans Multimed Comput, Commun Appl 19(5s):1–20
Xiang S, Qian D, Gao J et al (2023) Rethinking person re-identification via semantic-based pretraining[J]. ACM Trans Multimed Comput Commun Applic 20(3):1–17
Yi D, Lei Z, Liao S et al (2014) Learning face representation from scratch. arXiv preprint arXiv:1411.7923
Zheng T, Deng W (2018) Cross-pose lfw: A database for studying cross-pose face recognition in unconstrained environments. Beijing University of Posts and Telecommunications. Tech Rep 5(7)
Zheng T, Deng W, Hu J (2017) Cross-age lfw: A database for studying cross-age face recognition in unconstrained environments. arXiv preprint arXiv:1708.08197
Acknowledgements
This work was partially supported by the National Natural Science Foundation of China under Grant No.62301315, Startup Fund for Young Faculty at SJTU under Grant No.23X010501967 and Shanghai Municipal Health Commission Health Industry Clinical Research Special Project under Grant No.202340010. The authors would like to thank the anonymous reviewers for their valuable suggestions and constructive criticisms.
Funding
This work was partially supported by the National Natural Science Foundation of China under Grant No. 62301315, Startup Fund for Young Faculty at SJTU under Grant No. 23X010501967 and Shanghai Municipal Health Commission Health Industry Clinical Research Special Project under Grant No. 202340010.
Author information
Authors and Affiliations
Contributions
Suncheng Xiang and Hongwei Xu contributed conception and design of the study, as well as the experimental process and interpreted model results. Suncheng Xiang obtained funding for the project and provided clinical guidance. Hongwei Xu, Mingye Xie and Suncheng Xiang drafted the manuscript. All authors contributed to manuscript revision, read and approved the submitted version.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Ethics approval
Not Applicable. The datasets and the work do not contain personal or sensitive information, no ethical issue is concerned.
Consent to participate
The authors are fine that the work is submitted and published by Machine Learning Journal. There is no human study in this work, so this aspect is not applicable.
Consent for publication
The authors are fine that the work (including all content, data and images) is published by Machine Learning Journal.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Xiang, S., Xu, H., Xie, M. et al. SubFace: learning with softmax approximation for face recognition. Multimed Tools Appl (2024). https://doi.org/10.1007/s11042-024-19037-7
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11042-024-19037-7