Computer Vision -- ECCV 2020 : 16th European Conference, Glasgow, UK, August 23-28, 2020, Proceedings. Part X / Andrea Vedaldi, Horst Bischof, Thomas Brox, Jan-Michael Frahm (eds.).

Intro
Foreword
Preface
Organization
Contents
Part X
Discriminability Distillation in Group Representation Learning
1 Introduction
2 Related Work
3 Discriminability Distillation Learning
3.1 Formulation of Group Representation Learning
3.2 Formulation of Discriminability
3.3 Discriminability Distillation Learning
3.4 Feature Aggregation G
3.5 Advantage of Discriminability Distillation Learning
4 Experiments
4.1 Set-to-Set Face Recognition
4.2 Video-Based Person Re-identification
4.3 Action Recognition
5 Conclusion
References

Monocular Expressive Body Regression Through Body-Driven Attention
1 Introduction
2 Related Work
3 Method
3.1 3D Body Representation
3.2 Body-Driven Attention
3.3 Implementation Details
4 Experiments
4.1 Evaluation Datasets
4.2 Evaluation Metrics
4.3 Quantitative and Qualitative Experiments
5 Conclusion
References
Dual Adversarial Network: Toward Real-World Noise Removal and Noise Generation
1 Introduction
2 Related Work
2.1 Noise Removal
2.2 Noise Generation
3 Proposed Method
3.1 Two Factorizations of Joint Distribution

3.2 Dual Adversarial Model
3.3 Training Strategy
3.4 Network Architecture
4 Evaluation Metrics
5 Experimental Results
5.1 Experimental Settings
5.2 Results on SIDD Benchmark
5.3 Results on DND and Nam Benchmarks
6 Conclusion
References
Linguistic Structure Guided Context Modeling for Referring Image Segmentation
1 Introduction
2 Related Work
2.1 Semantic Segmentation
2.2 Referring Image Localization and Segmentation
2.3 Structural Context Modeling
3 Method
3.1 Multimodal Feature Extraction
3.2 Linguistic Structure Guided Context Modeling

3.3 Dual-Path Multi-Level Feature Fusion
4 Experiments
4.1 Experimental Setting
4.2 Comparison with State-of-the-Arts
4.3 Ablation Studies
5 Conclusion and Future Work
References
Federated Visual Classification with Real-World Data Distribution
1 Introduction
2 Related Work
3 Federated Visual Classification Problems
4 Datasets
4.1 INaturalist-User-120k and iNaturalist-Geo Splits
4.2 Landmarks-User-160k
5 Methods
5.1 Federated Averaging and Server Momentum
5.2 Importance Reweighted Client Objectives

5.3 Splitting Imbalanced Clients with Virtual Clients
6 Experiments
6.1 Classification Accuracy Vs Distribution Non-Identicalness
6.2 Importance Reweighting
6.3 Federated Virtual Clients
6.4 Federated Visual Classification Benchmarks
6.5 Implementation Details
7 Conclusions
References
Robust Re-Identification by Multiple Views Knowledge Distillation
1 Introduction
2 Related Works
3 Method
3.1 Teacher Network
3.2 Views Knowledge Distillation (VKD)
4 Experiments
4.1 Datasets
4.2 Self-distillation
4.3 Comparison with State-Of-The-Art