Proceedings of International Conference on Image, Vision and Intelligent Systems 2022 (ICIVIS 2022)

Description

This book is a collection of the papers accepted by the ICIVIS 2022―The International Conference on Image, Vision and Intelligent Systems, held on August 15–17, 2022, in Jinan, China. The topics focus but are not limited to image, vision and intelligent systems. Each part can be used as an excellent reference by industry practitioners, university faculties, research fellows and undergraduates as well as graduate students who need to build a knowledge base of the most current advances and state of practice in the topics covered by this conference proceedings. Preface Introduction Contents Image Denoising Supervision Based Generative Adversarial Networks for MRI Super-Resolution Reconstruction 1 Introduction 2 Related Work 2.1 Super-Resolution and Denoising of Single MR Image 2.2 Generative Adversarial Networks for SR Tasks 3 Method 3.1 Image Degradation Process 3.2 Generator 3.3 Discriminator and Loss Function 4 Experiments 4.1 Dataset and Implementation Details 4.2 Ablation Study 4.3 Compared with Other Methods 5 Conclusion References 3D Pot-Shaped Object Localization Based on RGB-Image and Point Clouds 1 Introduction 2 Related Work 2.1 3D Object Detection and Localization from RGB Images 2.2 3D Object Detection and Localization from Point Clouds 2.3 3D Object Localization Based on Image and Point Clouds 3 Methodology 3.1 Data Generation and Association 3.2 Object Feature Extraction from Images 3.3 Object Localization from Point Clouds 4 Experiments 4.1 Compare with Other 2D Detector 4.2 Effect of Different Voxel Size 4.3 Compare with Other Localization Methods 5 Conclusion References Edge Matching of Image with Large Viewpoint Variation Based on Relative Motion Direction of Edge Corners 1 Introduction 2 Edge Representation 3 Construction of Sequence of Relative Motion Direction of Edge Corners (RMDEC) of Feature Edge 4 Individual Edge Matching Based on RMDEC 5 Edge Matching of Images with Large Viewpoint Variation 6 Comparison Between RMDEC and State-of-the-Art Feature Matching Technology for Large Viewpoint Change Image 7 Conclusion References Research on Image Recognition in Complex Background Based on Improved Bilinear CNN Model 1 Introduction 2 Model Introduction 2.1 Bilinear CNN Model 2.2 Residual Structure Bilinear CNN Model 3 Model Training and Result Analysis 3.1 Dataset 3.2 Data Augmentation 3.3 Model Training 3.4 Analysis of Results 4 Summary References High Precision Odometer and Mapping Algorithm Based on Multi Lidar Fusion 1 Summary 2 Adaptive Calibration and Data Fusion 3 Data Preprocessing 3.1 Outlier Removal 3.2 Voxel Grid Filtering 3.3 Ground Point Filtration 4 Odometer and Mapping 4.1 Lidar Odometer 4.2 Mapping Update 5 Experimental Results and Analysis 6 Conclusion References Classification Model for Easily Confused Breast Ultrasound Image Based on Shape and Texture Features 1 Introduction 2 Related Work 3 Classification Model and Pipeline 4 Lesion Information Based on Morphology and Texture Information 4.1 Morphology Features 4.2 Texture Features 4.3 LMCL Loss Function Improvement 4.4 MT-LMCL Loss Function 5 Experiment 5.1 Data Set 5.2 Experiment Design 5.3 Result and Analysis 6 Conclusion References Automated Identification for High Voltage Isolation Switch via Generative Adversarial Network 1 Introduction 2 Related Work 3 Methodology 3.1 High Voltage Isolation Switch GAN 4 Experiment 4.1 Dataset and Evaluation Metrics 4.2 The Experimental Results of the Proposed Approach 5 Conclusion References Fundus Image Registration Based on Improved Feature Description 1 Introduction 2 Algorithm Implementation 2.1 Get Green Channel Image 2.2 Obtain Vessel Tree and Vessel Centerline 2.3 Vascular Feature Extraction 2.4 Feature Description 2.5 Feature Matching 3 Experimental Results 4 Final Match Result 5 Conclusion References Split Dynamic Target Counting Framework Based on Yolov5 +Deepsort Optimization 1 Introduction 2 Related Work 2.1 Target Detector 2.2 Messaging Protocols 2.3 Target Tracker 3 Framework 3.1 Target Detection Module 3.2 Signal Transmission System 3.3 Target Counting Function 4 Experiment 4.1 Dataset 4.2 Experimental Detail 4.3 Experimental Results 5 Conclusion References Clustering of Cervical Histopathology Images Based on Minimum Spanning Tree 1 Introduction 2 Materials and Methods 2.1 Histopathology Image Source 2.2 Research Method 2.3 k-Means Algorithm 2.4 Graph Theory 2.5 CHIC Method Based on MST 2.6 Manual Annotation and Results Evaluation 3 Result 3.1 Research Results and Analysis 4 Conclusion References Face Attribute Recognition Based on Multistage Adaptive Task Convolution Neural Network 1 Introduction 2 Methods 2.1 Data Processing 3 Method 3.1 Model 3.2 Training Process 3.3 Results 4 Conclusion References Generating Post-healing Images of Skin Diseases Based on an Adversarial Self-coding Generator 1 Introduction 2 Methods and Materials 2.1 Segmentation Module 2.2 Corrosion Module 2.3 VAE/GAN Based Fill-In Module 3 Experiments and Results 3.1 Segmentation 3.2 Different Stages of Erosion 3.3 The Generation of Post-healing Images 4 Conclusion References An Object Detection Method for Remote Sensing Images Based on DA-YOLO 1 Introduction 2 Related Work 3 Our Network 3.1 Quadruple Cropping 3.2 Network Architecture 4 Experiment and Analysis 4.1 Ablation Study 4.2 Comparison with Other Methods 4.3 Speed Experiment 4.4 Detection Result 5 Conclusion References Graph-Based Multi-granularity Person Re-identification 1 Introduction 2 Related Works 2.1 Component-Based Approaches 2.2 Graph-Based Approaches 3 Methodology 3.1 Graph Convolution 3.2 Loss Functions 4 Experiments 4.1 Data Set and Assessment Indicators 4.2 Implementation Details 4.3 Experimental Results and Analysis 4.4 Ablation Experiments 4.5 Parameter Analysis 5 Conclusion References Bendlets and Sparse Representation for Biological Image Inpainting 1 Introduction 2 The Bendlet Transform 3 Bendlet for Inpainting Biological Image 4 Experiments and Results 5 Conclusion References Convolutional Neural Network Design for Single Image Super-Resolution 1 Introduction 2 Related Work 2.1 Single Image Super-Resolution 2.2 Network Redundancy 3 The Proposed Method 3.1 Dissimilarity Between Features 3.2 Connection Pruning 3.3 Loss Function 3.4 Retrain Pipeline 4 Experiments 4.1 Datasets 4.2 Results 5 Conclusions References The Emergence and Evolution of Parallel Coordinates Plots 1 Introduction 2 The Emergence of Parallel Coordinates 3 The Evolution of Parallel Coordinates 3.1 The Improvement of Parallel Coordinates Plots by Data Mining Algorithm 3.2 The Improvement of Parallel Coordinates Plots by Dimensionality Reduction Algorithm 3.3 The Improvement of Parallel Coordinates Plots by Human-Computer Interaction Approach 3.4 The Improvement of Parallel Coordinates Plots by Other Visualization Methods 4 Conclusions and Future Directions References Compact Objects Extraction in Noisy Images 1 Introduction 2 Multi-threshold Object Selection Algorithm 2.1 Multi-threshold Object Selection Approach and Parameters of the Algorithm 2.2 Detection of Objects with Disk Shape 2.3 Object Selection by Area and Compactness 2.4 Decisive Statistics and Detection Efficiency 3 Object Distinguishing in Noisy Image 3.1 Objects Distinguishing in Gaussian Noise 3.2 Optimal Distinction of Objects in the Form of a Disk and a Square 3.3 Multi-threshold Algorithm for Object Selection and Distinction by Geometric Features 4 Object Extraction in Real Images 5 Conclusions References Micromodel and Analysis of COVID-19 Based on Hypergraph 1 Introduction 2 HC-SIRS Model 3 Model Analysis 4 Simulation Experiment and Results 4.1 Simulation Results of ‘Diamond Princess’ 4.2 Simulation Results of Other Cities 5 Concluding Remarks References Super-Resolution Reconstruction of License Plate Image Based on ESRGAN 1 Introduction 1.1 Background 1.2 Related Work 2 Method 2.1 ESRGAN 2.2 Network Structure 2.3 Loss Function 2.4 Evaluation Index 3 Experiments 3.1 Dataset 3.2 Experimental Result 4 Conclusion and Discussion References An Investigation into a Lung Nodule Segmentation Algorithm Based on Improved Swin-Unet 1 Methods of Research 1.1 Background 1.2 Swin-Unet Network Structure 1.3 Loss Function 1.4 Residual Block 2 Experiment and Evaluation 2.1 Dataset 2.2 Experimental Parameters and Results 3 Conclusion References Joint Learning Based on Discriminant Representation and Group Collaborative Measurement for Image Set Classification 1 Introduction 2 DRGCM 2.1 Local Multi Model Extraction 2.2 Prototype Point to Dictionary Representation 2.3 Test Point Set Dictionary Representation Learning 2.4 Mean Kernel Function Representation Classification 3 Experiments 4 Conclusion References UJN-SAR: A Large Dataset with Experimental Analysis for Water Body Segmentation from SAR Images 1 Introduction 2 Related Work 2.1 Water Body Segmentation Based on Traditional Methods 2.2 Water Body Segmentation Based on Deep Learning 2.3 Dataset Introduction 3 UJN-SAR Dataset 3.1 Image Source Introduction 3.2 UJN-SAR Introduction 4 Experimental Results and Analysis 4.1 Evaluation Metrics 4.2 Preprocessing Operation 4.3 Training Strategy 4.4 Result Analysis 5 Conclusion References Virtual Color Restoration for Ancient Architecture by Color Transfer 1 Introduction 2 Flow of the Proposed Method 3 Algorithm Description Flow of the Proposed Method 3.1 Color Space Conversion 3.2 Selection of Initial Clustering Centers 3.3 GMM Clustering 3.4 Nearest Region Matching 3.5 Color Transfer 4 Experiments and Discussion 4.1 Experiment 1 4.2 Experiment 2 5 Conclusions References Directional DCT Based Depth Image Coding for 3-D Video 1 Introduction 2 The Proposed Scheme 3 Experimental Results 4 Conclusions References PSAGNet: A Water Body Extraction Method for High Resolution Remote Sensing Images 1 Introduction 2 Related Works 2.1 Dilated Convolution 2.2 Attention Mechanism 2.3 Gated Convolution 3 Dataset and Method 3.1 Dataset and Data Pre-processing 3.2 PSAGNet Architecture 3.3 Gated Convolution Module 3.4 Pyramid Split Attention Module 3.5 Loss Function 3.6 Evaluation Metrics 4 Experiments 4.1 Experimental Setup 4.2 Experimental Result 4.3 Ablation Study 5 Conclusion References SEViT: A Large-Scale and Fine-Grained Plant Disease Classification Model Based on SE-ResNet and ViT 1 Introduction 2 Methods 2.1 SE-ResNet101 2.2 Vision Transformer 2.3 SEViT 3 Experiments 3.1 Dataset 3.2 Experiment Setting 3.3 Evaluation of SEViT 3.4 Classification Results of Plant Disease 3.5 Ablation Experiment 4 Conclusion References MSTRM: An Efficient Multi-scale Text Recognition Model 1 Introduction 2 Related Work 3 Methodology 3.1 Multi-object Rectification Network 3.2 Two-Branch Multi-scale Feature Fusion for ECA Attention 4 Experiments 4.1 Experimental Details 4.2 Evaluation Criteria 4.3 Analysis of Experimental Results 5 Conclusion References Recognition of Mushroom Using Deep Convolutional Neural Networks Based on Attention Mechanisms 1 Introduction 2 Materials and Methods 2.1 Image Dataset 2.2 Overall Process 2.3 Proposed Approach 2.4 Training Strategy 3 Results and Discussion 3.1 Training Result of Different Attention Modules 3.2 Training Result of Transformer-Based Model 4 Conclusion References SPPFU-Net: A Fusion Downsampling Network Focusing on Multiscale Features 1 Introduction 2 Related 3 Model Architecture 3.1 Spatial Pyramid Pooling Attention 3.2 Fusion Downsampling 4 Experiments and Results 4.1 Experimental Data 4.2 Evaluation Metrics 4.3 Experiments and Results Analysis 5 Conclusion References Cascade Detector Based on Multi-scale Features for Small Objects in Various Backgrounds of Remote Sensing 1 Introduction 2 Related Work 3 Method 3.1 Image Dehazing 3.2 Res2Net 3.3 PAFPN 3.4 Cascade R-CNN 4 Experiment and Results Discussion 4.1 Implementation Details 4.2 Evaluation Matrices 4.3 Results and Discussion 5 Conclusion References Machine Learning Classification of Alzheimer’s Disease Using Joint Features of Diffusion Tensor Imaging and Clinical Scales 1 Introduction 2 Materials and Methods 2.1 Materials 2.2 Methods 2.3 Experimental Setup 3 Results 3.1 Feature Selection 3.2 Performance Evaluation 4 Discussion 5 Conclusion References Research Progress of Improved Conventional Image Edge Detection Algorithm 1 Introduction 2 Traditional Edge Detection Methods 2.1 First-Order Differential-Based Edge Detection 2.2 Edge Detection Based on Second-Order Differentiation 2.3 Comparison of Traditional Edge Detection Operators 3 Improved Traditional Edge Detection Method 3.1 Improved Roberts Edge Detection Operator 3.2 Improved Prewitt Edge Detection Operator 3.3 Improved Sobel Edge Detection Operator 3.4 Improved Laplacian Edge Detection Operator 3.5 Improved LOG Edge Detection Operator 3.6 Improved DOG Edge Detection Operator 4 Conclusion References Analysis of Users’ Demand for Visual Images of Chinese Medicine Time-Honored Brands Based on the Kano Model 1 Introduction 2 Based on the KANO Model, the Research on the Demand for Chinese Time-Honored Pharmaceutical Brands 2.1 Introduction to KANO Model 2.2 Selection of User Experience Elements of Pharmaceutical Brands 2.3 Elements of User Demand Indicators of Pharmaceutical Brands 3 Demand Survey of Time-Honored Chinese Pharmaceutical Brands Based on the KANO Model 3.1 KANO Questionnaire Design 3.2 Analysis of KANO Questionnaire Survey Results 3.3 Demand Index Hierarchy Single Sort 3.4 Solution 4 Strategies for the Development of Time-Honored Chinese Medicine Brands 4.1 Brand Image of Time-Honored Chinese Medicine Brand Co-branded IP Image 4.2 Strengthen the Time and Interest of the Time-Honored Chinese Medicine Brand 4.3 Improve the Unity of the Brand Image of the Time-Honored Chinese Medicine Brand and Strengthen the Brand Memory 4.4 Enhance the Personalized Visual Experience of the Brand Image of the Time-Honored Chinese Medicine Brand 5 Conclusion References Brain Age Estimation by Convolutional Neural Network Based on T1-Weighted MRI and Gender Feature 1 Introduction 2 Materials and Methods 2.1 Materials 2.2 Models 2.3 Experimental Environment and Parameters Design 2.4 Evaluation Indicators 3 Results 4 Discussion 5 Conclusion References Axial Transformer U-Net for Cell Nuclei Segmentation in Gastrointestinal Cancer Pathological Image 1 Introduction 2 Related Work 2.1 Traditional Machine Learning-Based Methods 2.2 Convolutional Neural Network-Based Methods 2.3 Transformer Based Methods 3 Method 3.1 Framework Overview 3.2 Color Normalization 3.3 Gated Axial-Attention 3.4 Main Structure of the Network 3.5 Datasets 3.6 Performance of the Algorithm 4 Discussion 5 Conclusion References Dynamic Texture Segmentation Based on Improved Mixtures of Hidden Markov Model 1 Introduction 2 Improved Mixtures of Hidden Markov Model 2.1 Mixtures of Hidden Markov Model 2.2 Improved Mixtures of Hidden Markov Model 3 Experimental Results 4 Conclusion References Multi-scale Object Detection in Remote Sensing Images Based on Dual Attention Mechanism 1 Introduction 2 Related Work 2.1 Multi-size Object Detection 2.2 Multi-directional Object Detection 3 Our Network 3.1 Dual Attention Mechanism 3.2 Multi-scale Feature Fusion 3.3 Rotation Loss Function 4 Experiment 4.1 Dataset 4.2 Experimental Setup 4.3 Experimental Results 4.4 Ablation Experiment 5 Conclusion References Detection of Arbitrary Shaped License Plates Based on Key-Points Locations 1 Introduction 2 Related Work 3 The Proposed Approach 3.1 Detection Network Architecture 4 Experiments 4.1 Datasets 4.2 Quantitative Comparison 4.3 Analysis 5 Conclusion References Normative Prior Network for Anomaly Segmentation in Retinal OCT Images 1 Introduction 2 Method 2.1 Segmenting Retinal Regions with Net U 2.2 Learning Prior Normative Distribution with Net V 2.3 Anomaly Segmentation via Structural Similarity 3 Experiments 3.1 Datasets 3.2 Experimental Settings 3.3 Quantitative Analysis 3.4 Qualitative Analysis 3.5 Parameters Analysis 3.6 Ablation Study 4 Conclusion References Spatial-Spectral Siamese and Similarity Network for Hyperspectral Image Classification 1 Introduction 2 Proposed Method 2.1 Spatial-Spectral Siamese Network and Similarity Network 2.2 Supervised Clustering Tasks 2.3 Total Loss Function 3 Experimental Results 3.1 Data Description 3.2 Experimental Setting 3.3 Experimental Results 4 Conclusion References Mask Constraint Based Image Local Style Transfer Algorithm 1 Introduction 2 Image Style Transfer Based on CNN 3 An Image Local Style Transfer Algorithm Proposed in This Paper 3.1 Image Local Style Transfer Combined with Mask 3.2 Using Mask to Extract Target and Non Target Regions 4 New Definition of Local Style Transfer Content Loss Function 5 New Definition of Local Style Migration Style Loss Function 6 Experimental Results and Analysis 6.1 Experimental Environment 6.2 Datasets 6.3 Evaluation Method 6.4 Experimental Analysis 7 Conclusion References UJN-CD: A Large Scale Dateset for Change Detection with Comprehensive Analysis Based on Deep Learning 1 Introduction 2 Related Work 2.1 Technical Research 2.2 Existing Dataset 3 Dataset 3.1 Dataset Image Source 3.2 Introduction of UJN-CD Dataset 4 Experimental Results and Analysis 4.1 Evaluation Indicators 4.2 Training Strategy 4.3 Result Analysis 4.4 Backbone Network Modification 5 Conclusion References Aerial Image Dehazing Network Compression: Towards Micro-UAV’S Real-Time Online Inferencing 1 Introduction 2 Related Work 2.1 GAN Dehazing and GAN Compression 2.2 Knowledge Distillation 3 Methodology 3.1 Lightweight Residual Structure 3.2 Channel Pruning 3.3 Multi Granularity Knowledge Distillation Scheme 4 Experiments 4.1 Experimental Settings 4.2 Experimental Results 5 Conclusion References Research on Magnetic Resonance Imaging Technology Based on 23Na 1 Introduction 2 Principle of 23Na Magnetic Resonance Imaging 2.1 Detection Characteristic Factors of the Tested Body 2.2 Performance Factors of MRI Equipment 3 Key Technologies of 23Na MRI Equipment 3.1 Main Magnetic Field Based on 23Na Magnetic Resonance Imaging 3.2 RF Coil Based on 23Na Magnetic Resonance Imaging 3.3 Gradient Coil Based on 23Na Magnetic Resonance Imaging 3.4 Scan Sequence Based on 23Na Magnetic Resonance Imaging 3.5 Reconstruction Method Based on 23Na Magnetic Resonance Imaging 4 Application of 23Na Magnetic Resonance Imaging Technique 5 Conclusion References Research on the Detection Algorithm of Rural Road Surface Damage Based on Image Processing 1 Foreword 2 Image Processing Algorithm 2.1 VGG16 Model 2.2 The SSD Classification Model 2.3 SGD Algorithm 3 Implementation of the Rural Rural Road Damage by Detection Algorithm 3.1 Acquisition of Rural Rural Road Images 3.2 Preprocessing of the Image Data 4 Conclusion References Transmission Line Fault Location Method Based on Monocular Image and Lidar Information Fusion 1 Introduction 2 Transmission Line Fault Location Method Based on Monocular Image and Lidar 2.1 Data Federation Modeling 3 Test Verification 4 Conclusion References Image Super-Resolution Using Mixed Attention Module 1 Introduction 2 Related Work 2.1 SR Based on DL 2.2 Attention 2.3 SR Based on Attention 3 Method 3.1 Overall Framework 3.2 MAB 4 Experiment 4.1 Setting 4.2 Ablation Experiment 4.3 Model Analysis and Comparison 5 Conclusion References Quantitative Research on Deformation Angle of Arteriovenous Crossing Compression 1 Introduction 2 Vascular Segmentation Model 2.1 PF-Pooling 2.2 Evaluation Indicators 2.3 Segmentation Performance 2.4 Segmentation Results 3 Compression Angle Quantification 3.1 Double-Edge Separation 3.2 Quantification of Compression Degree at Arteriovenous Junction 3.3 Comparison of Results 4 Conclusion References Image-Based Self-attentive Multi-label Weather Classification Network 1 Introduction 2 Self-attention Based Multi-label Classification Network 2.1 Overall Structure of the Model 2.2 Multi Self-attention Mechanism [10] 2.3 Image Feature Extraction Network (CNN) 3 Experiment 3.1 Label Dependency Modeling 3.2 Classification Performance on Weather Datasets 4 Conclusion References Fire Detection Method Based on Infrared Image and Visible Image 1 Introduction 1.1 A Subsection Sample 2 Fire Detection Principle 2.1 Flame Detection Based on Threshold Segmentation 2.2 Smoke Detection Based on Deep Learning 3 Methods Used in This Article 3.1 Infrared Image Flame Detection 3.2 Smoke Detection Based on Improved Yolo Model 4 Experimental Analysis 4.1 Experimental Setup 4.2 Dataset 4.3 Experimental Result 4.4 Deploy Experimental Results References A Review of Fingerprint Image Enhancement Based on Gabor Filter 1 Introduction 2 Fingerprint Image Enhancement Based on Gabor Filter 3 Conclusion References Graph Spatial-Spectral Deep Subspace Clustering for Hyperspectral Image Classification 1 Introduction 2 Proposed Method 2.1 Patch-Level Features Extraction 2.2 Construction of Band Adjacency Matrix 2.3 Band-Level Features Extraction 2.4 Joint Feature Fusion Strategy 3 Experimental Results 3.1 Datasets and Implement Details 3.2 Classification Results 3.3 Parameters Analysis 3.4 Ablation Study 4 Conclusion References Research of Improved Deep Transfer Learning Algorithm in Image Classification 1 Introduction 2 Related Background 3 Algorithm Research 4 Experiments and Analysis 5 Conclusion References Automatic Liver Segmentation and 3D Imaging for CT Images Based on FANU-Net 1 Introduction 2 Related Works 2.1 3D Modeling 2.2 Image Segmentation 3 3D Reconstruction of Liver Based on Segmented Images 3.1 3D Modeling 3.2 Process Flow 4 Experimental Results 4.1 Database 4.2 Image Segmentation Results 4.3 3D Modeling Results 5 Conclusion References An Overview of Robust Secret Image Sharing 1 Introduction 2 Preliminaries 2.1 Introduction to Robust Secret Image Sharing 2.2 Robust Secret Image Sharing Metrics 2.3 Definition of Robust Secret Image Sharing 3 Robust Variants 3.1 Robust SIS Scheme Resist to Fake Shadows 3.2 Robust Secret Image Sharing Scheme Resist to Image Processing 4 Conclusion References Vision Implementation of Improved Yolov5-Based Backroom Violation Detection 1 Introduction 2 Related Work 3 Implementation of Backroom Staff Violation Detection 4 Analysis of Experimental Results 4.1 Experimental Platform 4.2 Experimental Platform 5 Conclusion 5.1 A Subsection Sample References VCSOD: A Video Conversion Scheme Based on Salient Object Detection Algorithm 1 Introduce 2 Related Work 3 Methodology 3.1 Salient Object Detection 3.2 Key Coordinate Points 3.3 System Construction 4 Experiment 4.1 Settings 4.2 Implementation 4.3 Analysis 5 Conclusion References Deep Convolutional Neural Network Based Cervical Cancer Exfoliated Cell Detection 1 Introduction 2 Related Work 3 Cervical Cell Classification Network Based on Deep Learning and Attention Mechanism 3.1 Data Preprocessing 3.2 Convolutional Neural Networks and Transfer Learning 3.3 Attention Mechanism 3.4 Evaluation Indicators 4 Experimental Results and Analysis 4.1 Database 4.2 Experimental Environment 4.3 Analysis of Experimental Results 5 Conclusion and Future Work References Intelligent Detection of Screw Looseness Fault in Running Gear of Electric Locomotive 1 Introduction 2 YOLOv3 3 Intelligent Detection of Screw Looseness 3.1 Screw Image Recognition 3.2 Evaluating Indicator 3.3 Screw Loosening Detection 4 Experiments and Analysis 4.1 Screw Dataset 4.2 Network Prediction 4.3 Looseness Determination 5 Conclusion References LMSA: Low-Relation Multi-head Self-attention Mechanism in Visual Transformer 1 Introduction 2 Related Work 2.1 Multi-head Self-attention Mechanism 2.2 Efficient Attention 2.3 Visual Transformer 3 Methodology 3.1 Low-Relation Multi-head Self-attention 3.2 Complexity Analysis 3.3 Data Efficiency Analysis 3.4 Comparison with Other Efficient Attention 4 Experiments 4.1 Image Classification 4.2 Semantic Segmentation 4.3 Ablation Study 5 Conclusion References Vehicle Detection Technology Based on HSV Color Model and Edge Detection 1 Introduction 2 Related Work 3 Main Methods 3.1 HSV Color Model 3.2 Edge Detection Operator 3.3 Inter-frame Difference Method 3.4 Binarization 4 System Design Process 5 Test Results and Analysis 6 Conclusion References Multi-receptive Field and Multi-stage Fusion Network for Infrared Small Target Detection 1 Introduction 2 Related Works 3 Method 3.1 Network Architecture 3.2 Multi-respective Field Module 3.3 Multi-stage Fusion Module 4 Experiments 4.1 Data Setting and Evaluation Metrics 4.2 Implementation Details and Comparison of the Results 4.3 Ablation Study 5 Conclusion References An Infrared Pedestrian Detection Method Based on Improved YOLOX 1 Introduction 2 YOLOX Network Structure 3 G-YOLOX Network Structure 3.1 Ghostnet 3.2 CBAM 3.3 G-YOLOX Network Structure 4 Analyze Experimental Results 4.1 Experimental Parameters 4.2 Model Performance Comparison 4.3 Actual Performance of the Model 5 Conclusion References Detection of Sea Surface Objects Based on Attention Mechanism and Feature Cross Aggregation 1 Introduction 2 Methods 2.1 Overview 2.2 Transformer Structure 2.3 CBAM Block 2.4 Feature Cross Module 2.5 Decoupled Head 3 Experimentations 3.1 Experimental Preparation 3.2 Experiment Result 4 Conclusion References Multiple Object Detection Based Semen Microscopic Video Analysis 1 Introduction 2 Basic Knowledge of the System 2.1 Knowledge in the Field of Semen Quality Diagnosis 2.2 Video Capture Environment 3 System Implementation and Achievement Display 3.1 Video Preprocessing 3.2 Image Preprocessing 3.3 kNN Classifier Design 3.4 Velocity Calculation and Sperm Classification 3.5 Visualization of Sperm Movement Path 4 Experimental Results and Analysis 4.1 Experimental Data and Environment 4.2 Experimental Results and Analysis 5 Conclusion References Face Expression Recognition with Occlusion Using Spatial Position Encoding and Attention Mechanisms 1 Introduction 2 Related Work 3 Method 3.1 Spatial Position Encoding Block 3.2 Parallel Attention Block 3.3 Feature Fusion Block and Loss Function 4 Experiments 4.1 Datasets 4.2 Implementation Details 4.3 Results of the Experiments 4.4 Ablation Analysis 4.5 Comparative Evaluation and Analysis of Experiments 5 Conclusion References Relay Contact Performance Detection Based on High Precision 3D Point Cloud Registration 1 Introduction 2 Relay Contact Surface Contact Performance Testing 2.1 Data Preprocessing 2.2 Model Matching 2.3 Defect Detection 3 Experiment 3.1 Experimental Platform 3.2 Contact Surface Inspection Process 3.3 Experimenta Conclusion 4 Conclusion References Garbage Classification Based on GC-Net 1 Introduction 2 Related Work 3 Background and Method 3.1 Datasets 3.2 Transfer Learning 3.3 PHAM 3.4 FReLU 3.5 Ensemble Learning 4 Experimental Result and Discussion 4.1 Experiment Setup 4.2 Experimental Result 5 Conclusion References Lightweight Object Detection Algorithm Based on Hybrid Pruning Strategy 1 Introduction 2 Our Approach 2.1 Multiscale Detection Algorithm Based on Hybrid Pruning Strategy 2.2 Attention-Based Fusion Module 2.3 Hybrid Pruning Strategy 3 Experiment and Analysis 3.1 Data Analysis and Preprocessing 3.2 Training Strategy 3.3 Evaluating Indicator 3.4 Ablation Experiment 3.5 Comparison with State-of-the-Art Methods 4 Conclusion References Driver State Monitoring System Based on YOLOv5 and Dlib 1 Introduction 2 System Modeling and Problem Formulation 2.1 Composition of the Driver State Monitoring System 2.2 Dataset Sources 2.3 Motivation of This Paper 3 Algorithm and Method 3.1 Structure Diagram of Driver State Detection Model 3.2 Construct the Driver State Detection Model 4 Experiment Result 5 Conclusion References Semi-supervised Person Re-identification Method Based on Voting Algorithm 1 Introduction 2 Related Work 2.1 Person Re-ID 2.2 Unsupervised Person Re-ID 2.3 Semi-supervised Person Re-ID 3 Method 3.1 Symbol Description 3.2 Supervised Part 3.3 Unsupervised Part 3.4 Cross-Domain Penalty 4 Experiment 4.1 Dataset and Implementation Details 4.2 Comparison 4.3 Ablation Experiment 5 Conclusion References Intelligent Systems From Motor to Battery: A Digital Twin Model of Electric Vehicles 1 Introduction 2 System Modeling and Problem Formulation 2.1 Battery Management System 2.2 Motivation Description 3 The Proposed Digital Twin Model 3.1 Algorithm Introduction 3.2 Method Introduction 4 Experiment Results 5 Conclusion References The SOC Prediction of Li-ion Batteries Based on Neural Network and Random Forest 1 Introduction 2 Problem Descriptions and Methodologies 2.1 Definition of Battery SOC 2.2 Random Forest 2.3 Neural Network 2.4 Genetic Algorithm 3 Prediction Model Training 3.1 Flow Chart 3.2 Data Acquisition 3.3 Data Preprocessing 3.4 Network Training 4 Experimental Results and Analysis 5 Conclusion References Design of Automatic Lane Change System Based on UN/ECE R79 1 Introduction 2 Automatic Lane Change System Scheme 2.1 Lane Change Intention Generation 2.2 Lane Change Feasibility Analysis 2.3 Lane Change Procedure 3 Experimental Results and Analysis 4 Conclusion References Ghost Module Network for Lightweight Weather Classification 1 The Introduction 2 Algorithm Based 2.1 The Depth of the Convolution 2.2 Ghost-Model 2.3 Channel Attention Mechanism 3 Data and Networks 3.1 Data Set 3.2 Network 3.3 Loss Function 4 The Experiment Part 5 Conclusion References A Multilevel Cognitive Diagnosis Model with Time Influencing Factors 1 Introduction 2 Related Work 2.1 Secondary Classification of Cognitive Diagnosis 2.2 Fuzzy Cognitive Diagnosis 2.3 Cognitive Diagnosis Combined with Deep Learning 2.4 Summary 3 Improved Cognitive Diagnosis Model 3.1 Typical Model of Cognitive Diagnosis 3.2 Improved Model Algorithm Flow 3.3 Mathematical Definition of Improved Model 4 Experiment 4.1 Data Set Introduction 4.2 Evaluating Indicator 4.3 Experimental Result 5 Concluding Remarks References Design and Implementation of Employee Recommendation System Based on Neural Graph Collaborative Filtering 1 Introduction 1.1 Research Background 1.2 The Traditional Recommendation Algorithms 1.3 Research Content 2 Materials and Methods 2.1 Datasets 2.2 Neural Graph Collaborative Filtering (NGCF) 2.3 System Design 3 Experiments and Analysis 3.1 Experiment Details 3.2 The Result 3.3 System Implementation 4 Conclusion References Digital-Twin-Driven Driving Range Prediction of Electric Vehicles 1 Introduction 2 System Modeling and Problem Formulation 2.1 Source of Data Sets Used in the Paper 2.2 Motivation for the Study 3 Algorithm and Method 3.1 Digital Twin Model Structure Diagram 3.2 Data Pre-processing 3.3 Regression Model 3.4 Performance Indicators 4 Experiment Result 5 Conclusion References Multi-scale Spatial Offset-Attention Network for 3D Point Clouds Classification 1 Introduction 2 Our Approach 2.1 Network Architecture 2.2 Graph Convolution 2.3 Spatial Offset-Attention 3 Experiments 3.1 Object Classification 3.2 Model Complexity 3.3 More Experiments on ModelNet40 4 Conclusion References A Multi-blockchain Based Reliable Noise Adding Method for Privacy Preservation in Cyber-Physical Systems 1 Introduction 2 System Model and Problem Setup 2.1 System Model 2.2 Problem Setup 3 RNA Method 3.1 Basic Idea 3.2 RNA Design 3.3 Algorithm Analysis 4 Performance Evaluation 4.1 Simulation Setup 4.2 Simulation Results 5 Conclusion References s Towards the Digital Twin Model of Li-Ion Batteries: State-of-Health (SoH) Prediction 1 Introduction 2 Related Work 3 Problem Analysis 3.1 Lithium-Ion Battery Health Status 3.2 Factors Affecting the Health Status of Lithium Ion Batteries 4 Model Establishment 4.1 Introduction to ElasticNet Model 4.2 The Idea of Applying ElasticNet Model to SoH Prediction 5 Experimental Results and Analysis 5.1 Experimental Data Collection 5.2 Feature Mining and Processing 5.3 Experiment Procedure 5.4 Result Analysis 6 Summarize References Research on Monthly Precipitation Prediction Based on Additive Ensemble Neural Network 1 Introduction 2 Mathematical Model of Additive Ensemble Neural Network 2.1 Basic Process of Additive Ensemble Neural Network 2.2 Determination of Weight of Additive Ensemble Neural Network 3 Empirical Analysis 3.1 Modeling Data 3.2 Model Performance Evaluation 3.3 Parameter Setting 3.4 Monthly Precipitation Prediction Model Based on Additive Ensemble Neural Network 4 Comparative Analysis of Two Models 4.1 Comparison of Results of Training Samples 4.2 Comparison of Results of Prediction Samples 4.3 Comparison of Fitting Accuracy 5 Conclusion References Forest Carbon Sequestration Estimation Based on Chapman-Richards and TOPSIS 1 Introduction 2 Model I: Carbon Sequestration Model 2.1 Model Preparation 2.2 Notations 2.3 Model Establishment 3 Model II: Forest Decision Model 3.1 Model Preparation 3.2 Model Establishment 3.3 Data Analysis and Suggestion 4 Model Application: Vienna Woods in Austria 4.1 Carbon Sequestration Model 4.2 Forest Decision Model 5 Conclusion References Adaptive Robust Command Filtered Control in Separate Meter-In and Separate Meter-Out Control System 1 Introduction 2 System Modeling 3 Controller Design 3.1 ARCFC Controller for Operating-Side System 3.2 ARCFC Design for Off-Side System 3.3 Stability Analysis 4 Experimental Results 5 Conclusion References Design and Research on Valve Train of Automobile Engine 1 Introduction 2 Mechanical Structure Design of Automatic Assembly Workstation for Automobile Valve Train 2.1 Quick Change Tool Hand Design and Pneumatic Schematic Diagram 2.2 Design of Tool Hand for Spring Lock and Its Pneumatic Schematic Diagram 2.3 Design of Tool Hand for Wheel Axle Installation and Valve Spring Pressing and Its Pneumatic Schematic Diagram 2.4 Design of Parts Silo 3 Design of Auto Engine Valve Train Automatic Assembly Workstation 3.1 Equipment Configuration 3.2 Design of Industrial Robot Fieldbus 3.3 Virtual Simulation of Overall Workstation 4 Conclusion References Prediction of Mini-mental State Examination Scores via Machine Learning for Alzheimer’s Disease 1 Introduction 2 Materials and Methods 2.1 Materials 2.2 Methods 3 Results 3.1 Results of M06 and M12 4 Discussion 5 Conclusion References Text Classification Model Based on Artificial Intelligence Algorithm 1 Introduction 2 Model 2.1 GCNN Layer 2.2 KNN-Attention Layer 3 Experimental Design and Evaluation 3.1 Data Set 3.2 Pretreatment 3.3 Experimental Results and Analysis 3.4 Parameter Analysis 3.5 Ablation Experiment 4 Conclusion References Information Aggregation and Fuzzy Decision Making Based on Vague Set Theory 1 Introduction 2 Vague Sets 3 Orders in the Vague Set System 4 Aggregation Function in Vague Set System 5 Fuzzy Decision Making Based on the Vague Set 6 Summary References Intelligent Cell Balancing of Li-Ion Batteries: A Particle Swarm Optimization Method 1 Introduction 2 System Modeling 2.1 Li-Ion Battery Modeling 2.2 Switch Resistor Circuit Modeling 3 Proposed Cell Balancing Approach 3.1 Particle Swarm Algorithm 3.2 Particle Swarm-Based Balancing 4 Simulation Results 4.1 Decentralized Cell Balancing Method 4.2 Proposed Cell Balancing Method 5 Conclusion References Photovoltaic Power Forecast Considering the Effect of Atmospheric Particulate Matter on Irradiance 1 Foreword 2 Correlation Analysis of Meteorological Factors 2.1 Data Acquisition 2.2 Relevance Analysis 3 Model Establishment 3.1 K-means Algorithm 3.2 RBF Neural Network 3.3 AOT Forecast Model 4 Estimation Model of Radiant Intensity of Inclined Plane 4.1 The Sum Amplitude Strength of the Sloped Surface 4.2 Model Example Analysis 5 PV Power Generation Power Forecast Model 5.1 Photoelectric Conversion Model 5.2 Validation of PV Power Generation Power Prediction Model 6 Summary References A Transition-Boosted Hashing for ECG Biometrics 1 Introduction 2 Related Works 3 Transition-Boosted Hashing 3.1 Inspiration 3.2 Transition-Boosted Hashing 3.3 Optimization 3.4 Online Search 4 Experimental Realization 4.1 Experiment Options 4.2 Result Analysis 5 Conclusion References Study on Trading Strategy Model of Financial Assets Based on Modified LSTM and Single-Objective Optimization Algorithm 1 Background 2 Reach Methods and Model 2.1 Forecast Method--ARIMA 2.2 Forecast Method--LSTM 2.3 Decision Method-- Single-Objective Optimization Model 3 Forecast of Asset Price 3.1 Forecast via ARIMA Model 3.2 Forecast via LSTM Model 3.3 The Best Trading Strategy 4 Tag References The Disambiguation Strategy of Knowledge Fusion for Domain Knowledge Graph 1 Introduction 2 Related Work 3 The Process of Knowledge Fusion 4 Evaluation Indicators and Results 5 Conclusion References Voice Assisted System for Garbage Sorting 1 Introduction 2 The Architecture of the Garbage Sorting System Based on Speech Recognition 3 Voice Recognition 3.1 Speech Recognition Forensics 3.2 Voice Recording 3.3 Audio Acquisition 3.4 Deepspeech2 Model for Speech Recognition 4 System Application Construction and Experiments 4.1 System Function Structure 4.2 Application Experiments 5 Conclusion References A Digital Twin Model of Smart Factory Production System 1 Introduction 2 Principle Description 3 Algorithms Used by the Digital Twin 3.1 Data Analysis and Preprocessing 3.2 Data Clustering 3.3 Data Classification 3.4 Regression 4 Experimental Results 4.1 Data Preprocessing 4.2 Data Clustering 4.3 Data Classification 4.4 Regression Prediction 4.5 Final Product 5 Conclusion References UAV Dynamic Trajectory Planning for Sudden Threat Detection 1 Introduction 2 Modeling of Theater Airspace Based on Voronoi Diagram 2.1 Problem Description 2.2 Airspace Modeling Under Threat 3 UAV Static Path Planning Based on Improved Ant Colony Algorithm [6] 3.1 Adjacency Matrix 3.2 Threat Cost 3.3 Time Benefit 3.4 Pheromone Accumulation Rule 3.5 Experimental Result 4 UAV Sudden Threat Track Planning 4.1 Sudden Threat 4.2 Design Scheme 4.3 Algorithm Example 5 Research Summary References Research on Hybrid Recommendation Algorithm Based on Incremental Clustering 1 Introduction 2 Related Work 2.1 Incremental Clustering Algorithm 2.2 KNN Algorithm 2.3 SVD Algorithm 3 Method 4 Experimental 4.1 Data Set 4.2 Evaluating Indicator 4.3 Experimental Result 4.4 Experimental Result 5 Conclusion References Optimal Sliding Mode Control of a Class of Affine Nonlinear Systems via Improved Successive Approximation Approach 1 Introduction 2 Preliminary Knowledge 3 Design 4 Simulations 5 Conclusion References Multi-objective Optimal Dispatch Method for Large-Scale Electric Vehicles Connected to Distribution Network 1 Introduction 2 Electric Vehicle Charging Behavior Characteristics 2.1 Daily Mileage 2.2 Charging Start Time 2.3 Charging Start SOC 2.4 Charging Time 2.5 Charging Power Model 2.6 Disordered Charging Load 3 Optimal Operation Strategy of Distribution Network 3.1 Goal Function 3.2 Restrictions 4 Algorithm Theory 4.1 NSGA-II Algorithm 4.2 AHP 5 Case Analysis 5.1 Study Settings 5.2 Study Settings 6 Summary References K-Anonymity-Based Privacy-Preserving and Efficient Location-Based Services for Internet of Vehicles Withstand Viterbi Attack 1 Introduction 2 Related Work 3 System Design 3.1 Overview 3.2 Improved K-Anonymity 3.3 Hybrid Caching Policy 4 Evaluation 4.1 Simulation Setup 4.2 Evaluation Results 5 Summary References Design and Implement Intelligent Discrimination of TCM Syndromes Based on Multi-label 1 Introduction 2 Relevant Theoretical Basis 2.1 The Theoretical Basis of Syndromes Differentiation 2.2 Theoretical Basis of Multi-label Learning 3 TCM Syndrome Intelligent Identification on Algorithm Based on Multi-label Learning 3.1 Key Problems to Be Solved 3.2 Algorithm Design and Implementation 4 Experimental Setup and Result Analysis 5 Conclusion References Research on Master-Slave Coordinated Control Method of Dual-Arm Robot 1 Introduction 2 Kinematics Analysis of Dual-Arm Robot 2.1 Forward Kinematics Model of Robot Arm 2.2 Inverse Kinematics Model of Robot Arm 3 Analysis of Dual-Arm Coordinated Control Method 3.1 Path Planning of End of the Master Arm 3.2 Constraints of the Slave Arm 4 Simulation of Dual-Arm Coordinated Control 4.1 Forward Kinematics Simulation of Single Arm 4.2 Inverse Kinematics Verification of Single Arm 4.3 Cooperative Control of Dual-Arm 5 Conclusions References Author Index