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

Intro
Foreword
Preface
Organization
Contents
Part XIII
Fashion Captioning: Towards Generating Accurate Descriptions with Semantic Rewards
1 Introduction
2 Related Work
3 The FAshion CAptioning Dataset
3.1 Data Collection, Labeling and Pre-Processing
3.2 Comparison with Other Datasets
4 Respecting Semantics for Fashion Captioning
4.1 Basic Problem Formulation
4.2 Attribute Embedding
4.3 Increasing the Accuracy of Captioning with Semantic Rewards
4.4 Joint Training of MLE and RL.
5 Experiments
5.1 Basic Setting
5.2 Performance Evaluations

6 Conclusion
References
Reducing Language Biases in Visual Question Answering with Visually-Grounded Question Encoder
1 Introduction
2 Related Works
3 Visually-Grounded Question Encoder (VGQE)
3.1 VGQE Cell
3.2 Using VGQE Cell to Encode the Question
3.3 Baseline VQA Architecture
4 Experiments and Results
4.1 Experimental Setup
4.2 Results
4.3 Performance of VGQE on Other Baselines
4.4 Performance of VGQE on the Standard VQAv2 Benchmark
5 Conclusion
References
Unsupervised Cross-Modal Alignment for Multi-person 3D Pose Estimation
1 Introduction

2 Related Work
3 Approaches
3.1 Architecture
3.2 Learning Cross-Modal Latent Space
3.3 Learning Beyond the Teacher Network
4 Experiments
4.1 Implementation Details
4.2 Ablation Studies
4.3 Datasets and Quantitative Evaluation
5 Discussion
6 Conclusion
References
Class-Incremental Domain Adaptation
1 Introduction
2 Background
3 Class-Incremental Domain Adaptation
3.1 Foresighted Source-Model Training
3.2 Class-Incremental DA on the Target Domain
4 Experiments
4.1 Discussion
5 Conclusion
References

Anti-bandit Neural Architecture Search for Model Defense
1 Introduction
2 Related Work
3 Anti-bandit Neural Architecture Search
3.1 Search Space
3.2 Adversarial Optimization for ABanditNAS
3.3 Anti-bandit
3.4 Anti-bandit Strategy for ABanditNAS
4 Experiments
4.1 Experiment Protocol
4.2 Results on Different Datasets
4.3 Ablation Study
5 Conclusion
References
Wavelet-Based Dual-Branch Network for Image Demoiréing
1 Introduction
2 Related Work
3 Our Method
3.1 Working in the Wavelet Domain
3.2 Dense Branch
3.3 Dilation Branch

3.4 Direction Perception Module
4 Experiments
4.1 Implementation Details
4.2 Datasets and State-of-the-Art Methods
4.3 Comparison with the State-of-the-Art
4.4 Ablation Studies
4.5 Extension to Deraining and Deraindrop
5 Conclusion
References
Low Light Video Enhancement Using Synthetic Data Produced with an Intermediate Domain Mapping
1 Introduction
2 Related Work
3 Learning the Low-Light Video RAW-to-RGB Mapping
3.1 Synthetic Data Generation Using an Intermediate Domain
3.2 Training Low-Light RAW-to-RGB Forward Models