We present a novel framework to exploit privileged information for recognition which is provided only during the training phase. Here, we focus on recognition task where images are provided as the main view and soft biometric traits (attributes) are provided as the privileged data (only available during training phase). We demonstrate that more discriminative feature space can be learned by enforcing a deep network to adjust adaptive margins between classes utilizing attributes. This tight constraint also effectively reduces the class imbalance inherent in the local data neighborhood, thus carving more balanced class boundaries locally and using feature space more efficiently. Extensive experiments are performed on five different datasets and the results show the superiority of our method compared to the state-of-the-art models in both tasks of face recognition and person re-identification.
Ensemble Soft-Margin Softmax Loss for Image Classification
Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);University of Chinese Academy of Sciences(中国科学院大学);Beihang University(北京航空航天大学);Tianjin University(天津大学);Macau University of Science and Technology(澳门科技大学)
CosFace: Large Margin Cosine Loss for Deep Face Recognition
Tencent AI Lab(腾讯人工智能实验室)
由于深度卷积神经网络(CNN)的进步,人脸识别取得了非凡的进展。人脸识别的核心任务包括人脸验证和身份识别,其中人脸特征鉴别是人脸识别的核心任务。然而,深度CNN的传统softmax损失通常缺乏区分能力。为了解决这个问题,最近已经提出了几种损失函数,例如中心损失、大裕度softmax损失和角度softmax损失。所有这些改进的损失都有一个相同的想法:最大化类间方差和最小化类内方差。在本文中,我们提出了一种新的损失函数,即大裕度余弦损失(LMCL),从不同的角度来实现这一想法。更具体地说,我们通过$L_2$对特征和权重向量进行归一化以去除径向变化,将softmax损失重新表示为余弦损失,在此基础上引入余弦裕度项以进一步最大化角空间中的决策裕度。结果,最小类内方差和最大类间方差是凭借归一化和余弦判决裕度最大化。我们将使用LMCL训练的模型称为CosFace。在最流行的公共领域人脸识别数据集上进行了广泛的实验评估,如MegaFace Challenge,Youtube Faces(YTF)和Labeled Face in the Wild(LFW)。我们在这些基准上实现了最先进的性能,这证实了我们所提出的方法的有效性。
JD AI Research(京东人工智能研究院);Institute of Automation, Chinese Academy of Science(中国科学院自动化研究所)
在人脸识别中,设计基于边缘的(例如,角、加性、加性角裕度)softmax损失函数在学习鉴别特征中起重要作用。然而,这些手工制作的探试法是次优的,因为它们需要花费很多精力来探索大的设计空间。最近,已经推导出用于损失函数搜索方法AM-LFS的AutoML,其利用强化学习在训练过程中搜索损失函数。但它的搜索空间复杂且不稳定,影响了它的优越性。本文首先分析了增强特征区分度的关键实际上是\textbf{how to reduce the softmax probability}。然后,我们设计了一个统一的公式,用于当前基于边际的softmax损失。因此,我们定义了一个新的搜索空间,并开发了一个奖励引导的搜索方法,自动获得最佳的候选人。各种人脸识别基准的实验结果表明,我们的方法的有效性超过国家的最先进的替代品。
Harbin Institute of Technology(哈尔滨工业大学);Peng Cheng Laboratory(鹏城实验室);King Abdullah University of Science and Technology(阿卜杜拉国王科技大学);Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学高瓴人工智能学院);Tsinghua University(清华大学)
DSAM: A Distance Shrinking with Angular Marginalizing Loss for High Performance Vehicle Re-identificatio
Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);National University of Singapore(新加坡国立大学);Macau University of Science and Technology(澳门科技大学)
SphereFace: Deep Hypersphere Embedding for Face Recognition
Georgia Institute of Technology(佐治亚理工学院);Carnegie Mellon University(卡内基梅隆大学);Sun Yat-Sen University(中山大学)
本文研究了开集协议下的深度人脸识别(FR)问题,其中理想的人脸特征在适当选择的度量空间下具有比最小类间距离更小的最大类内距离。然而,很少有现有的算法可以有效地实现这一标准。为此,我们提出了角度softmax(A-Softmax)损失,使卷积神经网络(CNN)能够学习角度判别特征。在几何上,A-Softmax损失可以被视为对超球面流形施加歧视性约束,这本质上匹配面也位于流形上的先验。此外,可以通过参数$m$定量地调整角度裕度的大小。我们进一步推导出具体的$m$来近似理想的特征标准。在Labeled Face in the Wild(LFW)、Youtube Faces(YTF)和MegaFace Challenge上的大量分析和实验表明,A-Softmax损失在FR任务中具有优越性。该守则也已公开提供。
AdaCos: Adaptively Scaling Cosine Logits for Effectively Learning Deep Face Representations
The Chinese University of Hong Kong(香港中文大学);SenseTime Research(商汤科技研究院);Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences(中国科学院深圳先进技术研究院)
Scalable Angular Discriminative Deep Metric Learning for Face Recognition
Nankai University(南开大学);Tianjin University(天津大学)
随着深度学习的发展,深度度量学习(Deep Metric Learning,DML)在人脸识别方面取得了很大的进步。具体来说,在训练过程中广泛使用的softmax损失往往会带来很大的类内变化,而特征归一化只在测试过程中使用,以计算对相似度。为了弥补这一差距,我们在训练步骤中将softmax损失中的特征和权重向量之间的类内余弦相似性设置为大于一个裕度,并从四个方面对其进行了扩展。首先,我们探讨了硬样本挖掘策略的效果。为了减轻调整超参数的人工劳动,提出了一种自适应的边缘更新策略。然后,一个规范化的版本,以充分利用余弦相似性约束。此外,我们加强了前一个约束,迫使类内余弦相似度大于平均类间余弦相似度,在指数特征投影空间中有一个保证金。在Labeled Face in the Wild(LFW)、Youtube Faces(YTF)和IARPA Janus Benchmark A(IJB-A)数据集上进行的大量实验表明,本文提出的方法优于主流的DML方法,接近最先进的性能。
Marginal Debiased Network for Fair Visual Recognition
School of Artificial Intelligence, Beijing Normal University(北京师范大学人工智能学院);School of Artificial Intelligence, Beijing University of Posts and Telecommunications(北京邮电大学人工智能学院)
NPCFace: Negative-Positive Collaborative Training for Large-scale Face Recognition
Shanghai University(上海大学);JD AI Research(京东AI研究院);National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Science(中国科学院自动化研究所模式识别国家实验室)
Neighborhood Watch: Representation Learning with Local-Margin Triplet Loss and Sampling Strategy for K-Nearest-Neighbor Image Classification
University of Arkansas(阿肯色大学);University of Washington(华盛顿大学);Washington University(华盛顿大学);Klinikum rechts der Isar(伊萨尔河右岸医院);Technical University of Munich(慕尼黑工业大学)
From Calibration to Refinement: Seeking Certainty via Probabilistic Evidence Propagation for Noisy-Label Person Re-Identification
School of Computer Science and Technology, Wuhan University of Science and Technology(武汉科技大学计算机科学与技术学院);Hubei Province Key Laboratory of Intelligent Information Processing and Real-Time Industrial System, Wuhan University of Science and Technology(湖北省智能信息处理与实时工业系统重点实验室);School of Computer Science, Wuhan University(武汉大学计算机学院);Department of Electrical Engineering and the Institute of Communications Engineering, National Tsing Hua University(清华大学电子工程系和通信工程研究所);Electronic and Optoelectronic System Research Laboratories, Industrial Technology Research Institute(工业技术研究院电子与光电子系统研究实验室)
Privileged Prior Information Distillation for Image Matting
Beijing University of Posts and Telecommunications(北京邮电大学);OPPO Research Institute(OPPO研究院);Xmotors;University of Maryland, Baltimore County(马里兰大学巴尔的摩县分校)
A Strong Balanced-Softmax Classifier-Retraining Baseline for Long-Tailed Recognition
CICATA Querétaro Instituto Politecnico Nacional(国立理工学院克雷塔罗中心研究与高级研究中心);Facultad de Informática Universidad Autónoma de Querétaro(克雷塔罗自治大学信息学院);Escuela Nacional de Estudios Superiores Universidad Nacional Autónoma de México(国立自治大学高等研究学校)
Sparsity-Inducing Divergence Losses for Biometric Verification
Athens University of Economics and Business(雅典经济与商业大学);Archimedes/Athena Research Center(阿基米德/雅典娜研究中心);Brno University of Technology(布尔诺理工大学);Omilia;National and Kapodistrian University of Athens(雅典大学)
Multi-view Contrastive Learning with Additive Margin for Adaptive Nasopharyngeal Carcinoma Radiotherapy Prediction
The Hong Kong Polytechnic University(香港理工大学);Research Institute for Smart Ageing, The Hong Kong Polytechnic University(香港理工大学智慧老龄研究学院);Nantong University(南通大学)
OPD-V: Visual On-Policy Self-Distillation with Modality Balance
National University of Singapore(新加坡国立大学);Ludwig Maximilian University of Munich(慕尼黑大学);Munich Center for Machine Learning(慕尼黑机器学习中心);Sun Yat-sen University(中山大学)
Prototype-Anchored Learning for Learning with Imperfect Annotations
Harbin Institute of Technology(哈尔滨工业大学);Peng Cheng Laboratory(鹏城实验室);King Abdullah University of Science and Technology(阿卜杜拉国王科技大学);Tsinghua University(清华大学)