Frontiers of Computer Vision : 29th International Workshop, IW-FCV 2023, Yeosu, South Korea, February 20–22, 2023, Revised Selected Papers

This book constitutes refereed proceedings of the 29th International Workshop on Frontiers of Computer Vision, IW-FCV 2023, held in Yeosu, South Korea in February 20–22, 2023. This workshop is an annual event that brings together researchers in the field of computer vision and artificial intelligence to share their research results. The workshop was started 29 years ago as a way to strengthen networking and share research results between Japanese and Korean researchers, and it has since grown in scope and influence, so from 2017, the workshop became an international event. The 13 full papers presented in this volume were thoroughly reviewed and selected from 72 submissions in 8 countries. The papers are dealing with the following topics:basic theories related to image processing, computer vision, image media, and human interface, as well as all research fields in applied fields such as autonomous vehicle driving, robot automation, and image content recognition. Recently, topics related to application of medical, bio, and entertainment field applying artificial intelligence, and more. Preface Organization Contents Multi-attributed Face Synthesis for One-Shot Deep Face Recognition 1 Introduction 2 Related Work 3 One-Shot Synthetic Face Recognition 3.1 Deep Synthetic Data Generation 3.2 Face Recognition 4 Evaluation Experiments 4.1 Datasets 4.2 Network Settings 4.3 Results 5 Limitations and Discussion References Efficient Multi-Receptive Pooling for Object Detection on Drone 1 Introduction 2 Related Work 3 The Proposed Method 3.1 The Backbone 3.2 Efficient Residual Bottleneck 3.3 Efficient Multi-Receptive Pooling 3.4 Loss Function 4 Implementation Details 5 Experimental Results 5.1 Evaluation on Datasets 6 Conclusion References Rough Target Region Extraction with Background Learning 1 Introduction 2 Related Work 3 Rough Target Region Extraction with Background Learning 3.1 Background Learning 3.2 Rough Target Region Extraction 4 Experiments 4.1 Experimental Setting 4.2 Evaluation of the Effectiveness of Decision-Boundary Induction 4.3 Comparison of Target Region Extraction Performance 4.4 Ablation Study 4.5 Limitation 5 Conclusions References Dynamic Circular Convolution for Image Classification 1 Introduction 2 Related Works 3 Methodology 3.1 Overall Architecture 3.2 Dynamic Circular Convolution 4 Experiments and Results 4.1 Experiments 4.2 Results 5 Conclusion References Classification of Lung and Colon Cancer Using Deep Learning Method 1 Introduction 2 Related Works 3 Datasets 3.1 LC25000 4 Methods 4.1 Cyclic Learning Rate 4.2 Proposed CNN Architecture 5 Result Analysis 5.1 Proposed Model Output 6 Conclusion References A Style-Based Caricature Generator 1 Introduction 2 Related Work 2.1 Caricature Creation 2.2 Style Transfer 3 Methodology 3.1 Face Caricature Creation 3.2 Style Transfer Generator 3.3 Implementation 4 Experiments 5 Conclusion References Attribute Auxiliary Clustering for Person Re-Identification 1 Introduction 2 Related Work 2.1 Unsupervised Person Re-ID Works 2.2 Attribute Auxiliary Person Re-ID 3 Methodology 3.1 USL Person Re-ID Pipeline 3.2 Attribute Auxiliary Clustering 4 Experiments 4.1 Datsets and Implementation 4.2 Comparison with State-of-the-Arts 4.3 Ablation Studies 5 Conclusions References YOLO5PKLot: A Parking Lot Detection Network Based on Improved YOLOv5 for Smart Parking Management System 1 Introduction 2 Related Work 2.1 Traditional-Based Method 2.2 Machine Learning-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 Texture Synthesis Based on Aesthetic Texture Perception Using CNN Style and Content Features 1 Introduction 2 Previous Research 3 Proposed Method 4 Quantification of Visual Impressions 4.1 Collection and Selection of Evaluation Terms 4.2 Subjective Evaluation Test 5 Extraction of Image Features 6 Construction of a Visual Impression Estimation Model 6.1 Construction of a Visual Impression Estimation Model 6.2 Results and Discussion 7 Texture Synthesis 7.1 Method 7.2 Results 8 Verification 8.1 Construction of Experimental Dataset 8.2 Effectiveness Verification Experiment 8.3 Results and Discussion 9 Conclusion References Robust Scene Text Detection Under Occlusion via Multi-scale Adaptive Deep Network 1 Introduction 2 Related Work 2.1 Scene Text Detection 2.2 Occluded Text Detection 3 Methodology 3.1 Transfer Learning Knowledge from Guided Grad-CAM++ Attention 3.2 Multi-scale Adaptive Deep Network 3.3 Loss Function 4 Experimental Results 4.1 Dataset 4.2 Implementation 4.3 Comparison Results 5 Conclusion References Classifying Breast Cancer Using Deep Convolutional Neural Network Method 1 Introduction 2 Related Works 3 Datasets 3.1 BreakHis 4 Methods 4.1 Preprocessing 4.2 Proposed CNN Architecture 5 Result Analysis 6 Conclusion References Front Cover Image Database of Japanese Manga and Typeface Estimation of Their Title 1 Introduction 2 Related Works 3 Proposed Method 3.1 Database of Front Cover Images 3.2 Typefaces of Fonts Used in Title 3.3 Label 3.4 Estimation Model 4 Experiments 4.1 Experimental Set-Up 4.2 Experimental Results 5 Conclusions Appendix References Human Face Detector with Gender Identification by Split-Based Inception Block and Regulated Attention Module 1 Introduction 2 Related Work 3 The Proposed Method 3.1 The Backbone 3.2 The Split-Based Inception Block 3.3 The Regulated Attention Module (RAM) 3.4 Classification Module 3.5 Face Detector 4 Experimental Settings 4.1 Dataset Pre-processing 4.2 Implementation Details 5 Results 5.1 Evaluation on Datasets 5.2 Ablation Study 5.3 Runtime Efficiency 5.4 Attention Modules Comparison 6 Conclusion References Author Index