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
03 Apr 2022
TL;DR: This study proposes two SKA variants where the first variant is applied in front of the ECAPA-TDNN model and the other is combined with the Res2net backbone block and demonstrates that they consistently improves the performance and are complementary when tested on three different evaluation protocols.
Abstract: The majority of recent state-of-the-art speaker verification architectures adopt multi-scale processing and frequency-channel attention mechanisms. Convolutional layers of these models typically have a fixed kernel size, e.g., 3 or 5. In this study, we further contribute to this line of research utilising a selective kernel attention (SKA) mechanism. The SKA mechanism allows each convolutional layer to adaptively select the kernel size in a data-driven fashion. It is based on an attention mechanism which exploits both frequency and channel domain. We first apply existing SKA module to our baseline. Then we propose two SKA variants where the first variant is applied in front of the ECAPA-TDNN model and the other is combined with the Res2net backbone block. Through extensive experiments, we demonstrate that our two proposed SKA variants consistently improves the performance and are complementary when tested on three different evaluation protocols.
4 citations
14 Apr 2019
TL;DR: This work proposes a self-paced deep learning model (SPDL) by introducing a negative \(l_1\)-norm regularizer for face recognition with label noise, inspired by curriculum learning, and achieves superior performance on LFW, MegaFace and YTF when there are different levels of label noise.
Abstract: Deep face recognition has achieved rapid development but still suffers from occlusions, illumination and pose variations, especially for face identification. The success of deep learning models in face recognition lies in large-scale high quality face data with accurate labels. However, in real-world applications, the collected data may be mixed with severe label noise, which significantly degrades the generalization ability of deep models. To alleviate the impact of label noise on face recognition, inspired by curriculum learning, we propose a self-paced deep learning model (SPDL) by introducing a negative \(l_1\)-norm regularizer for face recognition with label noise. During training, SPDL automatically evaluates the cleanness of samples in each batch and chooses cleaner samples for training while abandons the noisy samples. To demonstrate the effectiveness of SPDL, we use deep convolution neural network architectures for the task of robust face recognition. Experimental results show that our SPDL achieves superior performance on LFW, MegaFace and YTF when there are different levels of label noise.
4 citations
12 Oct 2020
TL;DR: In this article, the authors investigate how multimedia researchers can help tackle these problems to level the playing field for social media users by utilizing machine learning and multimedia content analysis techniques in a transparent manner and for the benefit of the users.
Abstract: Participation on social media platforms has many benefits but also poses substantial threats. Users often face an unintended loss of privacy, are bombarded with mis-/disinformation, or are trapped in filter bubbles due to over-personalized content. These threats are further exacerbated by the rise of hidden AI-driven algorithms working behind the scenes to shape users' thoughts, attitudes, and behaviour. We investigate how multimedia researchers can help tackle these problems to level the playing field for social media users. We perform a comprehensive survey of algorithmic threats on social media and use it as a lens to set a challenging but important research agenda for effective and real-time user nudging. We further implement a conceptual prototype and evaluate it with experts to supplement our research agenda. This paper calls for solutions that combat the algorithmic threats on social media by utilizing machine learning and multimedia content analysis techniques but in a transparent manner and for the benefit of the users.
4 citations
Posted Content•
TL;DR: A novel distillation method for lightweight neural network architectures that outperforms other known methods for the face recognition task on LFW, AgeDB-30 and Megaface datasets is proposed.
Abstract: The usage of convolutional neural networks (CNNs) in conjunction with a margin-based softmax approach demonstrates a state-of-the-art performance for the face recognition problem. Recently, lightweight neural network models trained with the margin-based softmax have been introduced for the face identification task for edge devices. In this paper, we propose a novel distillation method for lightweight neural network architectures that outperforms other known methods for the face recognition task on LFW, AgeDB-30 and Megaface datasets. The idea of the proposed method is to use class centers from the teacher network for the student network. Then the student network is trained to get the same angles between the class centers and the face embeddings, predicted by the teacher network.
4 citations
Cites background or methods from "ArcFace: Additive Angular Margin Lo..."
...For training procedure MS1MV2 [3] was used....
[...]
...Architecture Training method LFW % AgeDB-30 ResNet100 ArcFace[3] 99....
[...]
...We will consider ArcFace [3] as a special case of the margin-based softmax approach, since it gives the best performance among the margin-based methods:...
[...]
...They include Cosface [21], Sphereface [14], and ArcFace [3] approaches, which all can be described by the general formula:...
[...]
...21 (teacher) MobileFaceNet ArcFace[3] 99....
[...]
TL;DR: A simple Angular Triplet loss is introduced on the basis of analysis of different feature representations constrained by softmax loss and triplet loss, which can cooperate with softmax consistently and be seen as an effective baseline for vehicle re-identification task.
Abstract: Vehicle re-identification is becoming an increasingly important problem in modern intelligent transportation systems Substantial results have been achieved with methods based on deep metric learning Most of the previous works tend to design complicated neural network models or utilize extra information In this work, we introduce a simple Angular Triplet loss on the basis of analysis of different feature representations constrained by softmax loss and triplet loss A batch normalization layer with zero bias is adopted to pass through the embedded feature before loss calculation Then, triplet loss is calculated in cosine metric space instead of Euclidean space In this way, triplet loss can cooperate with softmax consistently By unifying the metric space of these two types of losses, the proposed method achieves 773% and 959% in rank-1 on VehicleID and VeRi-776 datasets, respectively With only global features utilized, the proposed model can be seen as an effective baseline for vehicle re-identification task
4 citations
Cites background from "ArcFace: Additive Angular Margin Lo..."
...A-Softmax loss [15], Additive Angular Margin loss [6] and Large Margin Cosine loss [30] can be seen as special linear classifier layers as they mainly act on the embedded features....
[...]
...[6] proposed a new angular margin loss by introducing a constant linear angular margin throughout the interval, which former works couldn’t guarantee....
[...]
References
More filters
27 Jun 2016
TL;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
Journal Article•
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
Proceedings Article•
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
Posted Content•
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