Mobile Networks and Management: 13th EAI International Conference, MONAMI 2023, Yingtan, China, October 27-29, 2023, Proceedings (Lecture Notes of the ... and Telecommunications Engineering, 559)

Preface
Organization
Contents
Wireless Communication and Networks
An Occlusion Signal-Processing Framework Based on UAV Sampling for Improving Rendering Quality of Views
	1 Introduction
	2 Related Work
		2.1 UAV Sampling and Rendering
		2.2 Light Field Sampling and Reconstruction
	3 A Framework for Occlusion Signal Based on UAV Sampling
		3.1 Aerial Light Field Parameterization
		3.2 Aerial Light Field Occlusion Signal Model
		3.3 Fourier Transform of the ALF Occlusion Signal Model
		3.4 ALF Occlusion Signal Sampling and Reconstruction
	4 Experiment
		4.1 Experimental Sampling Methods
		4.2 Real Scene Experiment
	5 Conclusion
	References
YOLOv5-LW: Lightweight UAV Object Detection Algorithm Based on YOLOv5
	1 Introduction
	2 Related Work
		2.1 Object Detection Algorithm
		2.2 Neck
	3 YOLOv5-LW
		3.1 Overview of YOLOv5
		3.2 YOLOv5-LW
	4 Experiments
		4.1 Experimental Environment
		4.2 Experimental Data
		4.3 Comparisons with the State-of-the-Art
		4.4 Ablation Studies
	References
Client Selection Method for Federated Learning in Multi-robot Collaborative Systems
	1 Introduction
	2 Related Work
	3 System Overview
		3.1 Task Scenario
		3.2 The Workflow of Fuzzy Logic
		3.3 The Workflow of Q-Learning
	4 A Method for Selecting Federated Learning Clients Based on Fuzzy Logic
		4.1 Fuzzy Set and Membership Functions
		4.2 Fuzzification and Defuzzification
	5 A Method for Selecting Federated Learning Clients Based on Q-Learning
		5.1 Design of Q-Table
		5.2 Design of Reward Function
		5.3 An Iterative Process of Q-Learning
	6 Experimental Result
		6.1 Experimental Dataset and Preprocessing
		6.2 Metrics
		6.3 A Method for Selecting Federated Learning Clients Based on Fuzzy Logic
		6.4 A Method for Selecting Federated Learning Clients Based on Q-Learning
	7 Conclusion
	References
Network Security and Blockchain
Cross-Chain Model of Notary Group Based on Verifiable Random Functions
	1 Introduction
	2 Blockchain Cross-Chain Technology
		2.1 Notary Scheme
		2.2 Sidechain/Relay Chain Technology
		2.3 Hash Locking Technology
		2.4 Distributed Private Key Control Technology
		2.5 The Current State of Improvements in the Notary Mechanism
	3 VRF-Based Notary Group Cross-Chain Model
		3.1 Cross-Chain Model of the Public Notary Group
		3.2 Verifiable Random Functions
	4 Analysis of Experimental Results
		4.1 Software and Hardware Test Environment
		4.2 Experiment and Result Analysis
	5 Conclusion
	References
Enhancing Cloud Data Integrity Verification Scheme with User Legitimacy Check
	1 Introduction
	2 Prerequisite
		2.1 Bilinear Pair
		2.2 Elliptic Curve Encryption Algorithm
		2.3 Homomorphic Hash Algorithm
		2.4 Verifiable Random Functions
	3 System Modeling
		3.1 Data Integrity Threat Model
	4 Data Integrity Protection Program
		4.1 Elliptic Curve Verifiable Random Function Design
		4.2 Data Upload
		4.3 User Legitimacy Check
		4.4 Data Integrity Validation
		4.5 Data Recovery
	5 Security Analysis
		5.1 User Legitimacy
		5.2 Data Integrity
		5.3 Preventing Replay Attacks
	6 Performance Analysis and Experimental Evaluation
		6.1 Fuel Consumption of Smart Contracts
		6.2 Comparison of Time Required to Generate Labels
		6.3 Comparison of Time Spent Verifying Integrity
		6.4 Error Finding
	7 Concluding Remarks
	References
Decoupled 2S-AGCN Human Behavior Recognition Based on New Partition Strategy
	1 Introduction
	2 Related Work
		2.1 ST-GCN Model
		2.2 2S-AGCN Model
	3 Decoupled 2S-AGCN Based on New Partition Strategy
		3.1 2S-AGCN with New Partition Strategy
		3.2 Decoupling GCN
	4 Experiment and Analysis
		4.1 Datasets
		4.2 Ablation Experiment
		4.3 Two-Stream Fusion Experiment
		4.4 Compared with Other Behavior Recognition Methods
	5 Conclusion
	References
Image Processing and Computer Vision
Human Behavior Recognition Algorithm Based on HD-C3D Model
	1 Introduction
	2 Human Behavior Recognition Based on C3D Model
	3 Improvements to This Paper
		3.1 Optimisation of Activation Functions
		3.2 Background Difference Method Optimisation of Dataset
		3.3 Image Scaling Improvements
	4 Experiments and Analysis of Results
		4.1 Improved Model HD-C3D
		4.2 Simulation Environment
		4.3 Analysis of Results
	5 Summary
	References
Inverse Pyramid Pooling Attention for Ultrasonic Image Signal Recognition
	1 Introduction
	2 Related Works
	3 Inverse Pyramid Pooling Attention Module
		3.1 Inverse Pyramid Pooling
		3.2 Inverse Pyramid Pooling ResNet
	4 Experiments and Analysis
		4.1 Experimental Preparation
		4.2 Comparison with the Most Advanced Methods
		4.3 Comparison of Model Improvement Effects
	5 Conclusion
	References
An Improved 4D Convolutional Neural Network for Light Field Reconstruction
	1 Introduction
	2 Proposed Method
		2.1 Image Enhancement Preprocessing
		2.2 HDAM
		2.3 GELU
	3 Experimental Results
		3.1 Real-World Scenes
		3.2 Synthetic Scenes
	4 Conclusion
	References
Fusing PSA to Improve YOLOv5s Detection algorithm for Electric Power Operation Wearable devices
	1 Introduction
	2 YOLOv5 Algorithm
	3 Improved YOLOv5s Algorithm
		3.1 Improvements of Backbone Module
		3.2 Improvements of Neck Module
		3.3 Improved YOLOv5s Network Structure
	4 Experimental Results and Analysis
		4.1 Data Set Production and Processing
		4.2 Training Network
		4.3 Comparison of Model Evaluation Between the Improved Algorithm and the Original Algorithm
	5 Conclusion
	References
Image Deblurring Using Fusion Transformer-Based Generative Adversarial Networks
	1 Introduction
	2 Related Work
		2.1 Design of the Generator
		2.2 Design of the Discriminator
		2.3 XT Moudle
		2.4 Data Normalization
		2.5 Loss Function
	3 Experimental Evaluation
		3.1 Datasets
		3.2 Comparative Experiment
		3.3 Subjective Deblurring Effect
		3.4 Ablation Experiment
	4 Conclusion
	References
Improved AODNet for Fast Image Dehazing
	1 Introduction
	2 AODNet Image Dehazing
		2.1 Atmospheric Scattering Model
		2.2 AODNet Dehazing
	3 Improved AODNet for Dehazing
		3.1 Feature Extraction Module
		3.2 Attention Mechanism Module
	4 Experiment and Analysis
		4.1 Experimental Scheme
		4.2 Comparative Analysis of Dehazing Performance
		4.3 Comparative Analysis of Dehazing Speed
	5 Summary
	References
Image Classification Algorithm for Graphite Ore Carbon Grade Based on Multi-scale Feature Fusion
	1 Introduction
	2 Multi-scale Feature Fusion Method for Graphite Ore Carbon Grade Image Classification
		2.1 Experimental Raw Data
		2.2 Graphite Ore Image Feature Extraction
	3 Experimental Results and Analysis
		3.1 Experimental Setup
		3.2 Network Model Ablation Experiments
		3.3 Comparative Experiments with Other Methods
	4 Conclusion
	References
Graphite Ore Grade Classification Algorithm Based on Multi-scale Fused Image Features
	1 Introduction
	2 Graphite Ore Image Classification and Recognition Method by Fusing Multi-scale Features
		2.1 Migration Learning and Graphite Ore Image Data Situation
		2.2 Graphite Ore Image Feature Extraction
	3 Experimental Analysis
		3.1 Experimental Setup
		3.2 Evaluation Metrics
		3.3 Network Model Effectiveness Ablation Experiment
		3.4 Quantitative Experimental Results Compared with Other Methods Are Presented in This Section
	4 Conclusion
	References
Improved WGAN for Image Generation Methods
	1 Introduction
	2 Related Content
		2.1 Generative Adversarial Networks
		2.2 WGAN
	3 Methods
		3.1 Residual Network
		3.2 Self-attention Mechanism
		3.3 Spectral Parametric Normalization
	4 Network Structure
		4.1 Generator Network
		4.2 Discriminator Network
	5 Experiment and Analysis
		5.1 Experimental Environment and Data Set
		5.2 Experimental Setup
		5.3 Experimental Results
		5.4 Comparison Experiments
		5.5 Ablation Experiments
	6 Conclusion
	References
Distributed Computing
Research on Multi-scale Pedestrian Attribute Recognition Based on Dual Self-attention Mechanism
	1 Introduction
	2 Related Work
	3 Methods
		3.1 Network Architecture
		3.2 Fusion Module
		3.3 Dual Self-Attention
	4 Experiments
		4.1 Datasets
		4.2 Evaluation Metrics
		4.3 Experimental Equipment and Settings
		4.4 Comparison Experiments
	5 Conclusion
	References
Improving Pedestrian Attribute Recognition with Dense Feature Pyramid and Mixed Pooling
	1 Introduction
	2 Methods
		2.1 Network Architecture
		2.2 Dense Feature Representation Module
		2.3 Feature Fusion Module
		2.4 Attribute Information Interaction Module
	3 Experiment Results and Analysis
		3.1 Data Set
		3.2 Evaluation Metrics
		3.3 Comparison Experiments
		3.4 Ablation Experiments
	4 Conclusion
	References
Double DQN Reinforcement Learning-Based Computational Offloading and Resource Allocation for MEC
	1 Introduction
	2 System Overview
		2.1 Network Module
		2.2 Computational Module
		2.3 Problem Formulation
	3 Algorithm Application
		3.1 Key Elements in RL
		3.2 DDQN Algorithm Application
	4 Simulation Result
	5 Conclusion
	References
Content Prediction for Proactive Tile-Based VR Video Streaming in Mobile Edge Caching System
	1 Introduction
	2 Model System
		2.1 Delay Model
		2.2 Computing Model
		2.3 Transmission Model
	3 Problem Formulation
		3.1 Performance Metric of QoE
		3.2 Joint Optimization
	4 Optimal Durations and Progressive Policy and VIE Online Caching Algorithm
		4.1 Optimal Durations
		4.2 Progressive Policy
		4.3 The VIE Online Caching Algorithm
	5 Simulation
	6 Conclusion
	References
Emerging Applications
Research on Latent Semantic Relationship Search Engine Based on Knowledge Graph
	1 Introduction
	2 Definition and Characteristics of Knowledge Graph
	3 Construction Process of Knowledge Graph
	4 The Architecture of Latent Semantic Search Engine Based on Knowledge Graph
	5 Summary
	References
Promoting Animation Synthesis Through Dual-Channel Fusion
	1 Introduction
	2 Relate Work
	3 First-Order Motion Model Animation Synthesis
		3.1 Improvement of Hourglass Neural Network in Motion Estimation
		3.2 Improved Generative Adversarial Network in FOMM
	4 Experiments
	5 Conclusion
	References
Author Index