ArcFace: Additive Angular Margin Loss for Deep Face Recognition
15 Jun 2019-pp 4690-4699
TL;DR: This paper presents arguably the most extensive experimental evaluation against all recent state-of-the-art face recognition methods on ten face recognition benchmarks, and shows that ArcFace consistently outperforms the state of the art and can be easily implemented with negligible computational overhead.
Abstract: One of the main challenges in feature learning using Deep Convolutional Neural Networks (DCNNs) for large-scale face recognition is the design of appropriate loss functions that can enhance the discriminative power. Centre loss penalises the distance between deep features and their corresponding class centres in the Euclidean space to achieve intra-class compactness. SphereFace assumes that the linear transformation matrix in the last fully connected layer can be used as a representation of the class centres in the angular space and therefore penalises the angles between deep features and their corresponding weights in a multiplicative way. Recently, a popular line of research is to incorporate margins in well-established loss functions in order to maximise face class separability. In this paper, we propose an Additive Angular Margin Loss (ArcFace) to obtain highly discriminative features for face recognition. The proposed ArcFace has a clear geometric interpretation due to its exact correspondence to geodesic distance on a hypersphere. We present arguably the most extensive experimental evaluation against all recent state-of-the-art face recognition methods on ten face recognition benchmarks which includes a new large-scale image database with trillions of pairs and a large-scale video dataset. We show that ArcFace consistently outperforms the state of the art and can be easily implemented with negligible computational overhead. To facilitate future research, the code has been made available.
Citations
More filters
•
01 Jan 2020
TL;DR: This thesis focuses on two learning tasks: learning to rank, and learning to classify, and reveals the importance of example weighting, by which a deep model focuses on more informative patterns, and pays less attention to non-informative and noisy ones during the learning process.
Abstract: In gradient-based optimisation, the derivative of the loss of an example can be interpreted as the example’s effect on the update of a model. Consequently, a derivative magnitude function can be considered to provide a weighting scheme from the viewpoint of example weighting. Therefore, example weighting is universal in deep learning. Partially arising from the recent work on the risky memorisation behaviours of deep neural networks (Arpit et al., 2017; Zhang et al., 2017b), example weighting becomes an active research filed (Chang et al., 2017; Toneva et al., 2019). Example weighting has ‘hard’ and ‘soft’ versions: (1) ‘hard’ weighting is well-known as sample selection or mining, i.e., binary weighting; (2) ‘soft’ weighting means example differentiation using a continuous importance score. In this thesis, we study how to learn more robust and discriminative representations using deep supervised learning. Technically, we propose example weighting for better optimisation and regularisation. Example weighting techniques differentiate and weight training data points according to a criteria, which varies in different scenarios. Example weighting improves the generalisation performance a lot, which is proved across multiple network architectures and learning tasks. We focus on two learning tasks in this thesis: learning to rank, and learning to classify. In both tasks, we reveal the importance of example weighting, by which a deep model focuses on more informative patterns, and pays less attention to non-informative (easy) and noisy (usually extremely hard) ones during the learning process. Therefore, example weighting is an important tool for guiding deep models to treat training samples differentially and learn meaningful patterns robustly and effectively. Furthermore, our study on example weighting helps us understand better about the training data and a model’s learning process. When a training dataset is clean, naively assigning higher weights to harder examples works well. However, when the dataset contains both meaningful and wrong information, a model learns meaningful patterns before fitting random errors. The challenge becomes how to differentiate trusted and error patterns as training progresses, and avoid fitting the error transformation. We demonstrate that example weighting is an effective approach for addressing this challenge. Additionally, we empirically justify the effectiveness of our proposed example weighting methods in other adverse cases:
3 citations
Cites methods from "ArcFace: Additive Angular Margin Lo..."
...There are many variants of this approach, including L2-Softmax (Ranjan et al., 2017), Large-margin Softmax (Liu et al., 2016a), Angular Softmax (Liu et al., 2017b), NormFace (Wang et al., 2017a), AM-Softmax (Wang et al., 2018a), CosFace (Wang et al., 2018b) and ArcFace (Deng et al., 2018)....
[...]
••
23 May 2022TL;DR: The authors proposed a multi-query multi-head attention (MQMHA) pooling and inter-top-K penalty method, which achieved state-of-the-art performance in all the public VoxCeleb test sets.
Abstract: This paper describes the multi-query multi-head attention (MQMHA) pooling and inter-topK penalty methods which were first proposed in our submitted system description for VoxCeleb speaker recognition challenge (VoxSRC) 2021. Most multi-head attention pooling mechanisms either attend to the whole feature through multiple heads or attend to several split parts of the whole feature. Our proposed MQMHA combines both these two mechanisms and gain more diversified information. The margin-based softmax loss functions are commonly adopted to obtain discriminative speaker representations. To further enhance the inter-class discriminability, we propose a method that adds an extra inter-topK penalty on some confused speakers. By adopting both the MQMHA and inter-topK penalty, we achieved state-of-the-art performance in all of the public VoxCeleb test sets.
3 citations
•
TL;DR: In this paper, a decision residual network (DRN) was proposed to capture uncertainty, enroll/test asymmetry and additional non-linear information in a 2nd-stage neural network (known as a decision network) for speaker recognition.
Abstract: Many neural network speaker recognition systems model each speaker using a fixed-dimensional embedding vector. These embeddings are generally compared using either linear or 2nd-order scoring and, until recently, do not handle utterance-specific uncertainty. In this work we propose scoring these representations in a way that can capture uncertainty, enroll/test asymmetry and additional non-linear information. This is achieved by incorporating a 2nd-stage neural network (known as a decision network) as part of an end-to-end training regimen. In particular, we propose the concept of decision residual networks which involves the use of a compact decision network to leverage cosine scores and to model the residual signal that's needed. Additionally, we present a modification to the generalized end-to-end softmax loss function to target the separation of same/different speaker scores. We observed significant performance gains for the two techniques.
3 citations
•
TL;DR: Zhang et al. as mentioned in this paper proposed a novel unsupervised federated face recognition approach (FedFR), which improves the performance in the target domain by iteratively aggregating knowledge from the source domain through federated learning.
Abstract: Unsupervised domain adaptation has been widely adopted to generalize models for unlabeled data in a target domain, given labeled data in a source domain, whose data distributions differ from the target domain. However, existing works are inapplicable to face recognition under privacy constraints because they require sharing sensitive face images between two domains. To address this problem, we propose a novel unsupervised federated face recognition approach (FedFR). FedFR improves the performance in the target domain by iteratively aggregating knowledge from the source domain through federated learning. It protects data privacy by transferring models instead of raw data between domains. Besides, we propose a new domain constraint loss (DCL) to regularize source domain training. DCL suppresses the data volume dominance of the source domain. We also enhance a hierarchical clustering algorithm to predict pseudo labels for the unlabeled target domain accurately. To this end, FedFR forms an end-to-end training pipeline: (1) pre-train in the source domain; (2) predict pseudo labels by clustering in the target domain; (3) conduct domain-constrained federated learning across two domains. Extensive experiments and analysis on two newly constructed benchmarks demonstrate the effectiveness of FedFR. It outperforms the baseline and classic methods in the target domain by over 4% on the more realistic benchmark. We believe that FedFR will shed light on applying federated learning to more computer vision tasks under privacy constraints.
3 citations
••
24 Oct 2022
TL;DR: This work investigates the gender bias of deep Face Recognition networks through a new post-processing methodology which transforms the deep embeddings of a pre-trained model to give more representation power to discriminated subgroups, and empirically observes that these hyperparameters are correlated with fairness metrics.
Abstract: In spite of the high performance and reliability of deep learning algorithms in a wide range of everyday applications, many investigations tend to show that a lot of models exhibit biases, discriminating against specific subgroups of the population (e.g. gender, ethnicity). This urges the practitioner to develop fair systems with a uniform/comparable performance across sensitive groups. In this work, we investigate the gender bias of deep Face Recognition networks. In order to measure this bias, we introduce two new metrics, $\mathrm{BFAR}$ and $\mathrm{BFRR}$, that better reflect the inherent deployment needs of Face Recognition systems. Motivated by geometric considerations, we mitigate gender bias through a new post-processing methodology which transforms the deep embeddings of a pre-trained model to give more representation power to discriminated subgroups. It consists in training a shallow neural network by minimizing a Fair von Mises-Fisher loss whose hyperparameters account for the intra-class variance of each gender. Interestingly, we empirically observe that these hyperparameters are correlated with our fairness metrics. In fact, extensive numerical experiments on a variety of datasets show that a careful selection significantly reduces gender bias.
3 citations
References
More filters
••
27 Jun 2016TL;DR: In this article, the authors proposed a residual learning framework to ease the training of networks that are substantially deeper than those used previously, which won the 1st place on the ILSVRC 2015 classification task.
Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.
123,388 citations
•
TL;DR: It is shown that dropout improves the performance of neural networks on supervised learning tasks in vision, speech recognition, document classification and computational biology, obtaining state-of-the-art results on many benchmark data sets.
Abstract: Deep neural nets with a large number of parameters are very powerful machine learning systems. However, overfitting is a serious problem in such networks. Large networks are also slow to use, making it difficult to deal with overfitting by combining the predictions of many different large neural nets at test time. Dropout is a technique for addressing this problem. The key idea is to randomly drop units (along with their connections) from the neural network during training. This prevents units from co-adapting too much. During training, dropout samples from an exponential number of different "thinned" networks. At test time, it is easy to approximate the effect of averaging the predictions of all these thinned networks by simply using a single unthinned network that has smaller weights. This significantly reduces overfitting and gives major improvements over other regularization methods. We show that dropout improves the performance of neural networks on supervised learning tasks in vision, speech recognition, document classification and computational biology, obtaining state-of-the-art results on many benchmark data sets.
33,597 citations
•
06 Jul 2015TL;DR: Applied to a state-of-the-art image classification model, Batch Normalization achieves the same accuracy with 14 times fewer training steps, and beats the original model by a significant margin.
Abstract: Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training, as the parameters of the previous layers change. This slows down the training by requiring lower learning rates and careful parameter initialization, and makes it notoriously hard to train models with saturating nonlinearities. We refer to this phenomenon as internal covariate shift, and address the problem by normalizing layer inputs. Our method draws its strength from making normalization a part of the model architecture and performing the normalization for each training mini-batch. Batch Normalization allows us to use much higher learning rates and be less careful about initialization, and in some cases eliminates the need for Dropout. Applied to a state-of-the-art image classification model, Batch Normalization achieves the same accuracy with 14 times fewer training steps, and beats the original model by a significant margin. Using an ensemble of batch-normalized networks, we improve upon the best published result on ImageNet classification: reaching 4.82% top-5 test error, exceeding the accuracy of human raters.
30,843 citations
28 Oct 2017
TL;DR: An automatic differentiation module of PyTorch is described — a library designed to enable rapid research on machine learning models that focuses on differentiation of purely imperative programs, with a focus on extensibility and low overhead.
Abstract: In this article, we describe an automatic differentiation module of PyTorch — a library designed to enable rapid research on machine learning models. It builds upon a few projects, most notably Lua Torch, Chainer, and HIPS Autograd [4], and provides a high performance environment with easy access to automatic differentiation of models executed on different devices (CPU and GPU). To make prototyping easier, PyTorch does not follow the symbolic approach used in many other deep learning frameworks, but focuses on differentiation of purely imperative programs, with a focus on extensibility and low overhead. Note that this preprint is a draft of certain sections from an upcoming paper covering all PyTorch features.
13,268 citations
•
TL;DR: The TensorFlow interface and an implementation of that interface that is built at Google are described, which has been used for conducting research and for deploying machine learning systems into production across more than a dozen areas of computer science and other fields.
Abstract: TensorFlow is an interface for expressing machine learning algorithms, and an
implementation for executing such algorithms. A computation expressed using
TensorFlow can be executed with little or no change on a wide variety of
heterogeneous systems, ranging from mobile devices such as phones and tablets
up to large-scale distributed systems of hundreds of machines and thousands of
computational devices such as GPU cards. The system is flexible and can be used
to express a wide variety of algorithms, including training and inference
algorithms for deep neural network models, and it has been used for conducting
research and for deploying machine learning systems into production across more
than a dozen areas of computer science and other fields, including speech
recognition, computer vision, robotics, information retrieval, natural language
processing, geographic information extraction, and computational drug
discovery. This paper describes the TensorFlow interface and an implementation
of that interface that we have built at Google. The TensorFlow API and a
reference implementation were released as an open-source package under the
Apache 2.0 license in November, 2015 and are available at www.tensorflow.org.
10,447 citations