Frontiers of Computer Vision. 30th International Workshop, IW-FCV 2024 Tokyo, Japan, February 19–21, 2024 Revised Selected Papers

Preface
Organization
Contents
Tackling Background Misclassification in Box-Supervised Segmentation: A Background Constraint Approach
	1 Introduction
	2 Related Work
		2.1 Condinst
		2.2 Boxinst
		2.3 Projection Loss
		2.4 Pairwise Loss
	3 Proposed Method
		3.1 Background Similarity
		3.2 Background Loss
		3.3 Foreground Loss
	4 Experiment
	5 Result
	6 Conclusion
	References
Clustering of Face Images in Video by Using Deep Learning
	1 Introduction
	2 Related Work
		2.1 Contrastive Learning
		2.2 Tsiam
	3 Our Approach
		3.1 Center Loss
		3.2 AutoEncoder
		3.3 Online Clustering
	4 Experiments
		4.1 Dataset
		4.2 Implementation Details
		4.3 Evaluation Metric
		4.4 Main Experiment
		4.5 Online Clustering
	5 Conclusion
	References
Exploring the Impact of Various Contrastive Learning Loss Functions on Unsupervised Domain Adaptation in Person Re-identification
	1 Introduction
	2 Related Work
		2.1 Unsupervised Person Re-identification
		2.2 Contrastive Learning in Person Re-identification
	3 Methodology
		3.1 UDA Person Re-identification Pipeline
		3.2 Unsupervised Training on the Target Domain Dataset
		3.3 Supervised Training on the Source Domain Dataset
	4 Experiments
		4.1 Dataset and Evaluation Metrics
		4.2 Implementation Details
		4.3 Discussions
		4.4 Comparison with State-of-the-art Methods
	5 Conclusion
	References
Automatic Measured Drawing Generation for Mokkan Using Deep Learning
	1 Introduction
	2 Related Work
	3 Proposed Method
		3.1 Preprocessing
		3.2 Image Conversion
		3.3 Postprocessing
		3.4 Dataset
		3.5 Training of Neural Network
	4 Evaluation
	5 Results and Discussion
	6 Conclusion
	References
Monocular Absolute 3D Human Pose Estimation with an Uncalibrated Fixed Camera
	1 Introduction
	2 Related Work
		2.1 Object Pose Estimation
		2.2 Human Pose Estimation
	3 Method
		3.1 Camera Calibration and Object Pose Estimation
		3.2 Absolute 3D Human Pose Estimation
		3.3 Visualization
	4 Evaluation
		4.1 User Study for Object Pose Estimation
		4.2 Quantitative Evaluation by Public Dataset
		4.3 Qualitative Evaluation by Original Dataset
		4.4 Discussion
	5 Conclusion
	References
Technical Skill Evaluation and Training Using Motion Curved Surface in Considered Velocity and Acceleration
	1 Introduction
	2 Motion Curved Surface Evaluation
		2.1 Creation and Display
		2.2 Calculation Method of Velocity and Acceleration Surface
		2.3 Velocity and Acceleration Surface Display
	3 Analyzation and Effectiveness Using Motion Curved Surface of Technical Skill
		3.1 Measurement of Technical Skill Motion
		3.2 Analyzation Using Motion Curved Surface
		3.3 Effectiveness of Technical Skill Training
	4 Conclusion
	References
A Benchmark for 3D Reconstruction with Semantic Completion in Dynamic Environments
	1 Introduction
	2 Related Work
		2.1 Semantic Scene Completion
		2.2 3D Human Motion Generation
	3 Dynamic 3D Scene Synthesis
		3.1 Static Scene Reconstruction with Complete Geometry
		3.2 Human Dynamics Simulation
		3.3 Scene-Motion Synthesis
	4 Experiments
		4.1 Experimental Settings
		4.2 Quantitative Results
		4.3 Qualitative Results
	5 Conclusion
	References
Framework for Measuring the Similarity of Visual and Semantic Structures in Sign Languages
	1 Introduction
	2 Theoretical Background
		2.1 Subspace Representation for Sign Language Video
		2.2 Video Features for Sign Subspaces
		2.3 Vector Representation of Words
	3 Proposed Method
		3.1 3D Visual and Semantic Maps on the Visual and Semantic Spaces
		3.2 Calculation of Communicability Metric
		3.3 Tsukuba New Signs Dataset
	4 Experimental Results
		4.1 Visualization of 3D Visual and Semantic Maps
		4.2 Visualization of Extracted Image Features
		4.3 Evaluation of the Communicability Metric
	5 Conclusion
	References
Human Facial Age Group Recognizer Using Assisted Bottleneck Transformer Encoder
	1 Introduction
	2 Related Work
	3 The Proposed Method
		3.1 The Feature Extraction Module
		3.2 The Assisted Bottleneck Transformer Encoder (ABTE)
		3.3 The Classification Module
	4 Implementation Settings
	5 Experiments and Results
		5.1 Evaluation on Datasets
		5.2 Model Analysis
		5.3 Runtime Efficiency
	6 Conclusion
	References
Efficient Detection Model Using Feature Maximizer Convolution for Edge Computing
	1 Introduction
	2 Related Work
		2.1 Edge Computing
		2.2 Efficient Feature Extraction
	3 Proposed Method
		3.1 Feature Maximizer Convolution
		3.2 Lightweight Strategy
	4 Experiment
		4.1 Dataset
		4.2 Evaluation Metric
		4.3 Experimental Setting
		4.4 Result
	5 Conclusion
	References
Spatial Attention Network with High Frequency Component for Facial Expression Recognition
	1 Introduction
	2 Related Work
	3 Proposed Method
	4 Experiment
		4.1 Dataset
		4.2 Experimental Setup
		4.3 Ablation Study
		4.4 Comparison
	5 Conclusion
	References
Minor Object Recognition from Drone Image Sequence
	1 Introduction
	2 Related Work
		2.1 Deep Neural Network-Based Method
		2.2 Convolutional Neural Network-Based Method
	3 Methodology
		3.1 Proposed Network Architecture
		3.2 Loss Function
	4 Experiments
		4.1 Dataset
		4.2 Experimental Setup
		4.3 Experimental Result
		4.4 Ablation Study
	5 Conclusion
	References
Author Index