Domain Adaptation for Visual Understanding

✍ Scribed by Richa Singh (editor), Mayank Vatsa (editor), Vishal M. Patel (editor), Nalini Ratha (editor)

• Publisher: Springer
• Year: 2020
• Tongue: English
• Leaves: 148
• Edition: 1st ed. 2020
• Category: Library

✦ Synopsis

This unique volume reviews the latest advances in domain adaptation in the training of machine learning algorithms for visual understanding, offering valuable insights from an international selection of experts in the field. The text presents a diverse selection of novel techniques, covering applications of object recognition, face recognition, and action and event recognition.

Topics and features: reviews the domain adaptation-based machine learning algorithms available for visual understanding, and provides a deep metric learning approach; introduces a novel unsupervised method for image-to-image translation, and a video segment retrieval model that utilizes ensemble learning; proposes a unique way to determine which dataset is most useful in the base training, in order to improve the transferability of deep neural networks; describes a quantitative method for estimating the discrepancy between the source and target data to enhance image classification performance; presents a technique for multi-modal fusion that enhances facial action recognition, and a framework for intuition learning in domain adaptation; examines an original interpolation-based approach to address the issue of tracking model degradation in correlation filter-based methods.

This authoritative work will serve as an invaluable reference for researchers and practitioners interested in machine learning-based visual recognition and understanding.

✦ Table of Contents

Preface
Contents
Contributors
Domain Adaptation for Visual Understanding
1 Introduction
2 Background: Transfer Learning
2.1 Notations and Definitions
2.2 Transfer Learning
3 Categories of Transfer Learning
3.1 Inductive Transfer Learning
3.2 Transductive Transfer Learning
3.3 Unsupervised Transfer Learning
4 Domain Adaptation
4.1 Homogeneous Versus Heterogeneous Domain Adaptation
4.2 Multistep Domain Adaptation
4.3 Related Areas
5 Summary
References
M-ADDA: Unsupervised Domain Adaptation with Deep Metric Learning
1 Introduction
2 Related Work
3 Proposed Approach: M-ADDA
4 Experiments
5 Conclusion
References
XGAN: Unsupervised Image-to-Image Translation for Many-to-Many Mappings
1 Introduction
2 Related Work
3 Proposed Model: XGAN
3.1 Architecture and Training Procedure
4 The CartoonSet Dataset
5 Experiments
6 Conclusions
References
Improving Transferability of Deep Neural Networks
1 Introduction
2 Related Work
3 Experimental Setup
4 Results and Discussion
4.1 Fine-tuning Last Layer
4.2 Fine-tuning Inner Layers
4.3 Graduated Fine-tuning of Inner Layers
5 Conclusion
References
Cross-Modality Video Segment Retrieval with Ensemble Learning
1 Introduction
2 Related Work
2.1 Vision Understanding
2.2 Language Understanding
2.3 Cross-Modal Understanding
3 Methods
3.1 Model Overview
3.2 Video Embedding
3.3 Language Embedding
3.4 Language-Based Ensemble
3.5 Late Fusion
4 Experiments
4.1 Experiment Setup
4.2 Implement Details
4.3 Result
5 Conclusion
References
On Minimum Discrepancy Estimation for Deep Domain Adaptation
1 Introduction
2 Related Works
3 Proposed Approach
3.1 Discussion
4 Experiments
4.1 Datasets
4.2 Experimental Setup
4.3 Results and Discussion
4.4 Visualization
5 Conclusion
References
Multi-modal Conditional Feature Enhancement for Facial Action Unit Recognition
1 Introduction
2 Related Work
3 Approach
3.1 Overview
3.2 Feature Extraction
3.3 Multi-modal Conditional Feature Enhancement (MCFE)
3.4 Training MCFE for AU Recognition
4 Experiments
4.1 Datasets
4.2 Settings
4.3 Results
5 Future Work
6 Conclusion
References
Intuition Learning
1 Introduction
1.1 Research Contributions
2 An Intuition Learning Algorithm
2.1 Adapting Feature Representation
2.2 Classification Using Reinforcement Learning
2.3 Design of Reward Function
2.4 Context-Dependent Addition Mechanism
3 Experimental Analysis
3.1 Dataset
3.2 Interpreting Intuition-Based Feature Subspace
3.3 Performance Analysis
4 Conclusion
References
Alleviating Tracking Model Degradation Using Interpolation-Based Progressive Updating
1 Introduction
2 Related Work
3 Alleviating Model Degradation Using Interpolation-Based Progressive Updating
3.1 Revisiting of CF-Based Tracking Method
3.2 Motion-Estimated Interpolation(MEINT) Based Progressive Updating
3.3 Process of Progressive Updating Model
3.4 Effect of Progressive Updating
4 Experiment
4.1 Dataset
4.2 Methods of Evaluation
4.3 Comparison Scenarios and Experimental Details
4.4 Experiment Result and Analysis
5 Conclusion
References
Index