Handbook of Face Recognition

• Autore: Li Stan Z
• Pubblicazione: Cham : , : Springer International Publishing AG, , 2024; ©2024
• Edizione: 3rd ed.
• Descrizione fisica: 1 online resource (473 pages)
• Disciplina: 006.42
• Altri autori: JainAnil K DengJiankang
• Nota di contenuto: Intro -- Foreword -- Preface to the Third Edition -- Preface to the Second Edition -- Acknowledgements -- Contents -- Contributors -- Part I Introduction and Fundamentals -- 1 Face Recognition Research and Development -- 1.1 Introduction -- 1.2 Processing Workflow -- 1.3 Advances in Deep Methods -- 1.3.1 Network Architectures -- 1.3.2 Loss Function -- 1.3.3 Face Recognition with GAN -- 1.3.4 Multi-task learning -- 1.4 Recent Development in Specific Face Recognition Tasks -- 1.4.1 Large-Scale Face Recognition -- 1.4.2 Cross-Factor Face Recognition -- 1.4.3 Masked Face Recognition -- 1.4.4 Heterogeneous FR -- 1.4.5 Low-Resolution Face Recognition -- 1.4.6 FR Under Atmospheric Turbulence -- 1.4.7 Face Recognition Against Adversarial Attacks -- 1.4.8 Fair Face Recognition -- 1.4.9 Video-Based Face Recognition -- 1.5 Databases -- 1.5.1 Overview of FR Datasets -- 1.5.2 Noisy Data -- 1.5.3 Data Imbalance -- 1.6 Other Related Topics -- 1.7 Conclusions -- 2 Convolutional Neural Networks and Architectures -- 2.1 Convolutional Neural Network Basics -- 2.1.1 Motivation: Idea Behind Convolutional Layer -- 2.1.2 Convolutional Layer: Concepts and Variants -- 2.1.3 CNN Example: AlexNet -- 2.1.4 Concept Modeling in CNNs -- 2.2 Convolutional Network Architectures -- 2.2.1 Going Deeper -- 2.2.2 Spatial Modeling -- 3 Generative Networks -- 3.1 Introduction -- 3.2 Variational Autoencoder -- 3.2.1 Vanilla VAE -- 3.2.2 Vector-Quantized Variational AutoEncoder -- 3.2.3 Variants of VAE and VQVAE -- 3.3 Generative Adversarial Networks -- 3.3.1 Overview -- 3.3.2 Architectures and Losses -- 3.4 Generative Model in Faces -- 3.4.1 Unconditional Generation -- 3.4.2 Conditional Generation -- 3.5 Evaluation of Generative Models -- 3.5.1 Evaluation Metrics on Sample Quality -- 3.5.2 Evaluation Metrics on Consistency -- 3.6 Summary -- Part II Face Processing Modules. ; 4 Face Detection -- 4.1 Introduction -- 4.2 The Challenges in Face Detection -- 4.3 Popular Face Detection Frameworks -- 4.3.1 Multi-stage and Two-Stage Face Detectors -- 4.3.2 One-Stage Face Detectors -- 4.4 Feature Extraction for Face Detection -- 4.4.1 Popular CNN Backbones -- 4.4.2 Toward Face Scale Invariance -- 4.4.3 Proposal Generation -- 4.5 Datasets and Evaluation -- 4.5.1 Datasets -- 4.5.2 Accuracy Evaluation Criterion -- 4.5.3 Results on Accuracy -- 4.6 Evaluation of the Computational Cost -- 4.6.1 FLOPs Versus AP in Multi-scale Test -- 4.6.2 FLOPs Versus AP in Single-Scale Test -- 4.6.3 FLOPs Versus Latency -- 4.7 Speed-Focusing Face Detectors -- 4.8 Conclusions and Discussions -- 5 Facial Landmark Localization -- 5.1 Introduction -- 5.2 Coordinate Regression -- 5.2.1 Coordinate Regression Framework -- 5.2.2 Network Architectures -- 5.2.3 Loss Functions -- 5.3 Heatmap Regression -- 5.3.1 Network Architectures -- 5.3.2 Loss Function -- 5.4 Training Strategies -- 5.4.1 Data Augmentation -- 5.4.2 Pose-Based Data Balancing -- 5.5 Landmark Localization in Specific Scenarios -- 5.5.1 3D Landmark Localization -- 5.5.2 Landmark Localization on Masked Face -- 5.5.3 Joint Face Detection and Landmark Localization -- 5.6 Evaluations of the State of the Arts -- 5.6.1 Datasets -- 5.6.2 Evaluation Metric -- 5.6.3 Comparison of the State of the Arts -- 5.7 Conclusion -- 6 Facial Attribute Analysis -- 6.1 Introduction -- 6.2 Facial Age Estimation -- 6.2.1 Regression-Based Methods -- 6.2.2 Classification-Based Methods -- 6.2.3 Ranking-Based Methods -- 6.2.4 Label Distribution Learning-Based Methods -- 6.3 Gender and Ethnicity Recognition -- 6.4 Multi-task Learning for Facial Attribute Estimation -- 6.5 Face Attribute Editing -- 6.5.1 Encoder-Decoder Structures -- 6.5.2 Image-to-Image Translation -- 6.5.3 Mask/Landmarks-Guided Architectures. ; 6.6 Recent Competitions -- 6.7 Datasets -- 6.7.1 Face Attribute Datesets -- 6.7.2 Deepfake Detection Datasets -- 6.8 Conclusions -- 7 Face Presentation Attack Detection -- 7.1 Introduction -- 7.2 Background -- 7.2.1 Face Presentation Attacks -- 7.2.2 Face PAD Pipeline in Face Recognition Systems -- 7.2.3 Camera Sensors for Face PAD -- 7.2.4 Face PAD Datasets -- 7.3 Methodology -- 7.3.1 Handcrafted Feature-Based Face PAD -- 7.3.2 Deep Learning-Based Face PAD -- 7.4 Experimental Results -- 7.4.1 Evaluation Metrics -- 7.4.2 Intra-dataset Testings -- 7.4.3 Cross-Dataset Testings -- 7.5 Applications -- 7.5.1 Online Identity Verification Scenario -- 7.5.2 Offline Payment Scenario -- 7.5.3 Surveillance Scenario -- 7.6 Conclusion and Future Challenge -- 8 Face Feature Embedding -- 8.1 Introduction -- 8.2 Loss Function -- 8.2.1 Softmax-Based Classification Loss Function -- 8.2.2 Metric Learning Loss Function -- 8.3 Network Architectures -- 8.3.1 General Networks -- 8.3.2 Specialized Networks -- 8.3.3 Mobile Networks -- 8.4 Large-Scale Training Datasets -- 8.5 Specific Face Recognition Topics -- 8.5.1 Long-Tail Learning -- 8.5.2 Cross-Variation Face Recognition -- 8.5.3 Noise-Robust Learning -- 8.5.4 Uncertainty Learning -- 8.6 Specific Loss Functions for Deep Face Recognition -- 8.6.1 Adaptive Curricular Learning Loss (CurricularFace) -- 8.6.2 Distribution Distillation Loss (DDL) -- 8.6.3 Sphere Confidence Face (SCF) -- 8.7 Conclusions -- 9 Video-Based Face Recognition -- 9.1 Introduction -- 9.2 Related Work -- 9.2.1 Pre Deep Learning Methods -- 9.2.2 Deep Learning Based Methods -- 9.3 Method -- 9.3.1 Face/Fiducial Detection -- 9.3.2 Deep Feature Representation -- 9.3.3 Face Association -- 9.3.4 Model Learning: Deep Subspace Representation -- 9.3.5 Matching: Subspace-to-Subspace Similarity for Videos -- 9.4 Experiments -- 9.4.1 Datasets. ; 9.4.2 Implementation Details -- 9.4.3 Evaluation Results -- 9.4.4 Cross-Spectral Video Face Verification -- 9.4.5 Discussions -- 9.5 Concluding Remarks -- 10 Face Recognition with Synthetic Data -- 10.1 Introduction -- 10.2 Related Work -- 10.3 Method -- 10.3.1 Deep Face Recognition -- 10.3.2 SynFace Versus RealFace -- 10.3.3 SynFace with Identity Mixup -- 10.3.4 SynFace with Domain Mixup -- 10.4 Experiments -- 10.4.1 Datasets -- 10.4.2 Implementation Details -- 10.4.3 Long-Tailed Face Recognition -- 10.4.4 Effectiveness of 147Depth148 and 147Width148 -- 10.4.5 Impacts of Different Facial Attributes -- 10.5 Conclusion -- 11 Uncertainty-Aware Face Recognition -- 11.1 Introduction -- 11.2 Background: Uncertainty-Aware Deep Learning -- 11.3 Probabilistic Face Embeddings (PFE) -- 11.3.1 Limitations of Deterministic Embeddings -- 11.3.2 Contidional Gaussian Distributions as Probabilistic Embeddings -- 11.3.3 Face Matching with PFEs -- 11.3.4 Template Feature Fusion with PFEs -- 11.3.5 Learning Uncertainty -- 11.3.6 Experimentals -- 11.3.7 Qualitative Analysis -- 11.3.8 Risk-Controlled Face Recognition -- 11.4 Learning Representations with Data Uncertainty -- 11.4.1 Classification-Based DUL -- 11.4.2 Regression-Based DUL -- 11.4.3 Experiments -- 11.5 Non-Gaussian Probabilistic Embeddings -- 11.5.1 r-radius von-Mises Fisher Distribution -- 11.5.2 Spherical Confidence Face (SCF) -- 11.5.3 Feature Comparison -- 11.5.4 Feature Pooling -- 11.5.5 Experiments -- 11.6 Summary -- 12 Reducing Bias in Face Recognition -- 12.1 The Bias in Face Recognition -- 12.2 Fairness Learning and De-biasing Algorithms -- 12.3 Problem Definition -- 12.4 Jointly De-biasing Face Recognition and Demographic Attribute Estimation -- 12.4.1 Adversarial Learning and Disentangled Representation -- 12.4.2 Methodology -- 12.4.3 Experiments. ; 12.5 Mitigating Face Recognition Bias via Group Adaptive Classifier -- 12.5.1 Adaptive Neural Networks -- 12.5.2 Methodology -- 12.5.3 Experiments -- 12.6 Demographic Estimation -- 12.7 Conclusion -- 13 Adversarial Attacks on Face Recognition -- 13.1 Introduction -- 13.2 Threat Model -- 13.2.1 Adversary's Goals -- 13.2.2 Adversary's Capabilities -- 13.2.3 Adversary's Knowledge -- 13.3 Digital Adversarial Attacks -- 13.3.1 White-Box Attacks -- 13.3.2 Transfer-Based Black-Box Attacks -- 13.3.3 Query-Based Black-Box Attacks -- 13.3.4 Universal Adversarial Attacks -- 13.4 Physical Adversarial Attacks -- 13.4.1 Patch-Based Physical Attacks -- 13.4.2 Light-Based Physical Attacks -- 13.5 Adversarial Defense for Face Recognition -- 13.5.1 Input Transformation -- 13.5.2 Adversarial Training -- 13.6 Positive Applications of Adversarial Attacks -- 13.7 Discussion -- 14 Heterogeneous Face Recognition -- 14.1 Introduction -- 14.1.1 Literature Review -- 14.2 Feature Descriptor-Based HFR -- 14.2.1 Graphical Representation -- 14.2.2 Similarity Metric -- 14.3 Face Synthesis-Based HFR -- 14.3.1 Deep Patch Representation Extraction -- 14.3.2 Candidate Patch Selection -- 14.3.3 Deep Graphical Representation Learning -- 14.4 Common Space-Based HFR -- 14.4.1 Network Architecture -- 14.4.2 Loss Function -- 14.5 Experiments -- 14.5.1 Databases -- 14.5.2 Heterogeneous Face Recognition Results -- 14.5.3 Heterogeneous Face Synthesis Results -- 14.6 Conclusion -- 15 3D Face Recognition -- 15.1 Introduction -- 15.2 Fast and Light Manifold CNN-based 3D FER -- 15.2.1 Method -- 15.2.2 Experiments -- 15.3 Low-Quality Depth Image Based 3D FR -- 15.3.1 Lock3DFace: A Large-Scale Database of Low-Cost Kinect 3D Faces -- 15.3.2 Led3D: A Lightweight and Efficient Deep Model to Low-cost 3D FR -- 15.3.3 Experiments -- 15.4 Nonlinear 3D Morphable Face Model -- 15.4.1 Method. ; 15.4.2 Experiments.
• Titolo autorizzato: Handbook of Face Recognition
• ISBN: 3-031-43567-2
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Lingua di pubblicazione: Inglese
• Record Nr.: 9910799228903321
• Lo trovi qui: Univ. Federico II