Marginal Loss for Deep Face Recognition

By Jiankang Deng, Stefanos P. Zafeiriou. 2017

224 citations recorded by the source corpus.

Convolutional neural networks have significantly boosted the performance of face recognition in recent years due to its high capacity in learning discriminative features. In order to enhance the discriminative power of the deeply learned features, we propose a new supervision signal named marginal loss for deep face recognition. Specifically, the marginal loss simultaneously minimises the intra-class variances as well as maximises the inter-class distances by focusing on the marginal samples. With the joint supervision of softmax loss and marginal loss, we can easily train a robust CNNs to obtain more discriminative deep features. Extensive experiments on several relevant face recognition benchmarks, Labelled Faces in the Wild (LFW), YouTube Faces (YTF), Cross-Age Celebrity Dataset (CACD), Age Database (AgeDB) and MegaFace Challenge, prove the effectiveness of the proposed marginal loss.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

Unified Contrastive Learning in Image-Text-Label Space (2022, methodology): “With this goal, numerous works have pushed the image recognition performance from different directions, such as data scale from MNIST [36] to ImageNet1K [10], model architectures from convolutional neural networks (CNNs) [23, 26, 34, 35, 40, 52, 53] to Transformers [15, 43, 57, 61, 64, 67, 72], and ”
Sub-center ArcFace: Boosting Face Recognition by Large-Scale Noisy Web Faces (2020, methodology): “Even though remarkable advances have been achieved by the margin-based soft-max methods [8,15,32,31,5,39], they all need to be trained on well-annotated clean datasets [30,5], which require intensive human efforts.”
Striking the Right Balance With Uncertainty (2019, unknown): “DeepFace [48] 97.35 91.4 FaceNet [42] 99.63 95.4 Web-scale [49] 98.37 - DeepID2+ [47] 99.47 93.2 Baidu [29] 99.13 - Center Face [55] 99.28 94.9 Marginal Loss [13] 99.48 96.0 Noisy Softmax [8] 99.18 94.9”
Deep Imbalanced Learning for Face Recognition and Attribute Prediction (2018, methodology): “Other works adopt ideas from metric learning, and combine softmax with contrastive [41], [42], [43], center loss [44] or marginal loss [45] for improvements.”
AgeDB: The First Manually Collected, In-the-Wild Age Database (2017, methodology): “Table 3: Age-invariant face verification in AgeDB utilizing the Centre Loss [50] and Marginal Loss [7] methods.”
ArcFace: Additive Angular Margin Loss for Deep Face Recognition (2018, background): “Several variants [36, 6, 43, 15, 35, 33, 4, 32, 25] have been proposed to enhance the discriminative power of the softmax loss.”
The Devil of Face Recognition is in the Noise (2018, methodology): “A common method is to query celebrities’ images by their names on search engines, and subsequently clean the labels with automatic or semi-automatic approaches [15,11,4].”
Global-Local GCN: Large-Scale Label Noise Cleansing for Face Recognition (2020, unknown): “CASIA [54] 97.11 92.93 Asian [1] 94.90 91.21 IMDb-Face [43] 97.87 96.26 VGGFace2 [12] 98.00 95.54 MS1M-IBUG [16] 98.25 97.53 MS1M-V2 [15] 98.48 98.35 MC - MegaFace2 [36] 98”
Unequal-Training for Deep Face Recognition With Long-Tailed Noisy Data (2019, methodology): “Another euclidean metric learning methods are based on classiﬁcation, such as Centerloss [39], Rangeloss [44], and Marginal loss [8].”
One-shot Face Recognition by Promoting Underrepresented Classes (2017, methodology): “We have not reproduced all these cost function design methods [33, 7, 29, 4] with our training dataset due to practical reasons.”
Deep Face Recognition: A Survey (2018, background): “[39] selected the farthest intra-class samples and the nearest inter-class samples to compute a margin loss.”
Past, Present, and Future of Face Recognition: A Review (2020, background): “activation [143] and various loss functions [15, 124-126, 130, 134, 136, 138, 141-142, 145, 148-152], the”
WebFace260M: A Benchmark Unveiling the Power of Million-Scale Deep Face Recognition (2021, methodology): “After steps of iteration, our fully automatic cleaning strategy provides purified data for model training, outperforming semi-automatic methods used in [14, 12, 1].”