Additive Margin Softmax for Face Verification

By Feng Wang, Jian Cheng, Weiyang Liu, Haijun Liu. 2018

1,380 citations recorded by the source corpus.

In this letter, we propose a conceptually simple and intuitive learning objective function, i.e., additive margin softmax, for face verification. In general, face verification tasks can be viewed as metric learning problems, even though lots of face verification models are trained in classification schemes. It is possible when a large-margin strategy is introduced into the classification model to encourage intraclass variance minimization. As one alternative, angular softmax has been proposed to incorporate the margin. In this letter, we introduce another kind of margin to the softmax loss function, which is more intuitive and interpretable. Experiments on LFW and MegaFace show that our algorithm performs better when the evaluation criteria are designed for very low false alarm rate.

Subfields: Computer Vision and Pattern Recognition

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

Citations in the corpus

Improving Neural Language Modeling via Adversarial Training (2019, background): “Direct Diversity Regularization There has been a body of literature on increasing the robustness by directly adding various forms of diversity-enforcing penalty functions (e.g., Elsayed et al., 2018; Xie et al., 2017; Liu et al., 2016; 2017; Chen et al., 2017; Wang et al., 2018).”
Continual Learning by Modeling Intra-Class Variation (2022, methodology): “The intuition of why MOCA can well regularize the representation space also comes from the series of works in large-margin softmax (Liu et al., 2016; 2017a;b; 2022; Wang et al., 2018a;b; Deng et al., 2019).”
Deep Unknown Intent Detection with Margin Loss (2019, methodology): “DNN such as BiLSTM (Goo et al., 2018; Wang et al., 2018c) has demonstrated the ability to learn high-level semantic features of intents.”
Noise Robust TTS for Low Resource Speakers using Pre-trained Model and Speech Enhancement (2020, methodology): “To extract more discriminative speaker embedding, additive margin softmax [17] is adopted to train the speaker recognition model.”
Learning towards Minimum Hyperspherical Energy (2018, background): “Current state-of-the-art face recognition methods [49, 28, 51, 6, 31] usually only focus on compressing the intra-class features, which makes MHE a potentially useful tool in order to further improve these face recognition methods.”
Learning Imbalanced Datasets with Label-Distribution-Aware Margin Loss (2019, methodology): “Recently, Large-Margin Softmax [37], Angular Softmax [38], and Additive Margin Softmax [53] have been proposed to minimize intra-class variation in predictions and enlarge the inter-class margin by incorporating the idea of angular margin.”
Orthogonal Over-Parameterized Training (2020, background): “[54, 52, 72, 10, 71, 47, 50] propose to learn representations on a hypersphere and show that the angular information, in contrast to magnitude information, preserves the most semantic meaning.”
ArcFace: Additive Angular Margin Loss for Deep Face Recognition (2018, methodology): “In SphereFace [15, 16], ArcFace, and CosFace [35, 33], three different kinds of margin penalty are proposed, e.”
Long-tail learning via logit adjustment (2020, background): “See also Liu et al. [2016, 2017], Wang et al. [2018], Khan et al. [2019] for similar weighting of negatives in the softmax.”
ELF: An Early-Exiting Framework for Long-Tailed Classification (2020, background): “tackle class imbalance by establishing margins depending upon the class size [23, 24, 25].”
Regularizing Neural Networks via Minimizing Hyperspherical Energy (2019, background): “The importance of regularizing angular information is also discussed in [31, 32, 33, 34, 35, 36, 37, 38, 39, 40].”
Speaker recognition based on deep learning: An overview (2020, neutral): “More information can be found in [129, 136, 137, 138].”
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.”
Review of Image Classification Algorithms Based on Convolutional Neural Networks (2021, background): “There are also many improved versions based on it, such as center-loss [82], LSoftmax [83], A-Softmax [84], AM-Softmax [85], PEDCC-loss [86], etc.”