10977nam 2200505 450 991050261680332120220627203005.03-030-87361-7(CKB)4100000012037952(MiAaPQ)EBC6737967(Au-PeEL)EBL6737967(OCoLC)1272994084(PPN)258051795(EXLCZ)99410000001203795220220627d2021 uy 0engurcnu||||||||txtrdacontentcrdamediacrrdacarrierImage and graphicsPart III 11th International Conference, ICIG 2021, Haikou, China, August 6-8, 2021, proceedings /Yuxin Peng [and five others], editorsCham, Switzerland :Springer,[2021]©20211 online resource (840 pages)Lecture notes in computer science : image processing, computer vision, pattern recognition, and graphics ;Volume 128903-030-87360-9 Includes bibliographical references and index.Intro -- Preface -- Organization -- Contents - Part III -- 3D Computer Vision -- Efficient Depth-Included Residual Refinement Network for RGB-D Saliency Detection -- 1 Introduction -- 2 Related Work -- 2.1 RGB Salient Object Detection -- 2.2 RGB-D Salient Object Detection -- 3 The Proposed Network -- 3.1 Architecture Overview -- 3.2 Depth Correction Module -- 3.3 Multi-scales Localization Module -- 3.4 Residual Refinement Module -- 4 Experiments -- 4.1 Experimental Setup -- 4.2 Compared with the State-of-the-arts -- 4.3 Ablation Study -- 5 Conclusion -- References -- Learning Cross-Domain Descriptors for 2D-3D Matching with Hard Triplet Loss and Spatial Transformer Network -- 1 Introduction -- 2 Related Works -- 2.1 2D Feature Descriptors -- 2.2 3D Feature Descriptors -- 2.3 Deep Similarity Learning Networks -- 2.4 2D-3D Cross-Domain Feature Descriptors -- 3 Methodology -- 3.1 Network Architecture -- 3.2 Spatial Transformer Network -- 3.3 Hard Triplet Loss -- 4 Experiments -- 4.1 Dataset -- 4.2 2D-3D Retrieval Task -- 4.3 3D Global Registration -- 4.4 Ablation Study -- 5 Conclusion -- References -- A Stereo Matching Method for Three-Dimensional Eye Localization of Autostereoscopic Display -- 1 Introduction -- 2 Autostereoscopic Display and Eye Localization -- 3 Stereo Matching and 3D Localization -- 3.1 The Principle of Binocular Stereo Vision -- 3.2 Image Matching Algorithm -- 4 An Eye Stereo Matching Algorithm Based on ZNCC -- 4.1 Algorithm Overview -- 4.2 Improvement Based on the Operation Logic of Matching Algorithm -- 4.3 Optimization Based on the Application Scenarios of Stereoscopic Display -- 5 Experiment and Discussion -- 6 Conclusion -- References -- Scaling Invariant Harmonic Wave Kernel Signature for 3D Point Cloud Similarity -- 1 Introduction -- 2 Related Works -- 2.1 Contribution -- 3 Fundamentals and Pipeline.3.1 3D Point Cloud Laplace-Beltrami Operator -- 3.2 Pipeline -- 4 Scaling Invariant Harmonic Wave Kernel Signature -- 4.1 Wave Kernel Signature -- 4.2 Scaling Invariant Harmonic Wave Kernel Signature -- 4.3 Invariance of the Harmonic Wave Kernel Signature -- 4.4 3D Point Cloud Models Similarity Measure -- 5 Experiments -- 5.1 Effectiveness of the SIHWKS -- 5.2 Robutness of the SIHWKS and Similarity Results -- 6 Conclusion -- References -- PST-NET: Point Cloud Sampling via Point-Based Transformer -- 1 Introduction -- 2 Related Work -- 2.1 Point Cloud Sampling -- 2.2 Point-Based Deep Learning -- 3 Network -- 3.1 Data Augmentation -- 3.2 Transformer -- 3.3 Loss -- 4 Experiments -- 4.1 ModelNet40 -- 4.2 Implementation Details -- 4.3 Controlled Experiments of Relation Function -- 4.4 Classification -- 5 Conclusion -- References -- Bird Keypoint Detection via Exploiting 2D Texture and 3D Geometric Features -- 1 Introduction -- 2 Related Work -- 2.1 Traditional Methods -- 2.2 Deep Learning Methods -- 2.3 Motivation of Our Method -- 3 Proposed Method -- 3.1 2D Branch -- 3.2 3D Branch -- 3.3 Fusion Module -- 4 Experiments -- 4.1 Implementation Details -- 4.2 Dataset and Evaluation Metrics -- 4.3 Comparison with State-of-the-arts -- 4.4 Ablation Study -- 5 Conclusion -- References -- Rotation Aware 3D Point Cloud Vehicle Detection -- 1 Introduction -- 2 Related Work -- 2.1 Point-Based Methods -- 2.2 Voxel-Based Methods -- 3 Method -- 3.1 Backbone and Detection Networks -- 3.2 Auxiliary Network -- 3.3 Loss Function -- 4 Experiment -- 4.1 Implementation Details -- 4.2 Evaluation on KITTI Validation Set -- 4.3 Ablation Studies -- 5 Conclusions -- References -- Improved 3D Morphable Model for Facial Action Unit Synthesis -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 3D Morphable Model -- 3.2 Improved 3DMM for AU Synthesis (AU-3DMM).4 Experiments -- 5 Conclusion -- References -- Semantic Guided Multi-directional Mixed-Color 3D Printing -- 1 Introduction -- 2 The System -- 2.1 The Hardware -- 2.2 The Software -- 3 Color Mixing Scheme -- 3.1 Extruder with Multiple Inputs -- 3.2 Color Mixing Control -- 3.3 Removing the Material with Transition Color -- 4 Semantic Guided Model Segmentation -- 4.1 User Assigned Model Semantics -- 4.2 Print Sequence Planning -- 4.3 Subdivision of Components -- 5 Experiment -- 5.1 Effectiveness of Color Switching Scheme -- 5.2 Some Examples -- 6 Conclusion -- References -- KeypointNet: Ranking Point Cloud for Convolution Neural Network -- 1 Introduction -- 2 Method -- 2.1 KeypointNet -- 2.2 KeypointNet Instantiation for CNN -- 2.3 Sampling Layer -- 3 Experimental Results -- 3.1 Comparison on Classification -- 3.2 Comparison on Segmentation -- 4 Conclusion -- References -- Geometric Context Sensitive Loss and Its Application for Nonrigid Structure from Motion -- 1 Introduction -- 2 Related Work -- 3 The Proposed GCS Loss Function -- 3.1 Revisiting the Mean Square Error Loss -- 3.2 The Proposed Geometric Context Sensitive Loss -- 4 Applications of the GCS Loss in NRSFM -- 4.1 Basic Formulation -- 4.2 Leverage GCS Loss -- 4.3 Optimization -- 5 Experiments -- 5.1 Datasets -- 5.2 Evaluation Metrics -- 5.3 Implementation Details -- 5.4 Baselines -- 5.5 Results and Discussions -- 6 Conclusions -- References -- 3D Reconstruction from Single-View Image Using Feature Selection -- 1 Introduction -- 2 Related Work -- 2.1 Voxel Grid -- 2.2 Other Representation of 3D Object -- 3 The Proposed Module -- 3.1 Point Selection in Inference and Training -- 3.2 Point-Wise Features and MLP Layer Head -- 4 Experiments -- 4.1 Datasets, Metrics and Loss Function -- 4.2 Implementation Details -- 4.3 Reconstruction of Synthetic Images -- 4.4 Space Complexity -- 5 Conclusions.References -- Computational Photography -- Adaptive Steganography Based on Image Edge Enhancement and Automatic Distortion Learning -- 1 Introduction -- 2 Related Work -- 2.1 Distortion Minimization Framework -- 2.2 Automatic Distortion Learning Steganography Methods Using GAN -- 3 The Proposed Framework -- 3.1 The Overall Architecture -- 3.2 Image Edge Enhancement -- 3.3 The Architecture of the Generator G -- 3.4 The Training Strategy -- 4 Experimental Results and Analysis -- 4.1 Experiment Setups -- 4.2 The Selection of the Image Edge Enhancement Methods -- 4.3 Comparison of Different Generators -- 4.4 Comparison with State-of-the-Art Methods -- 4.5 The Practical Application -- 5 Conclusion -- References -- No-Reference Image Quality Assessment via Broad Learning System -- 1 Introduction -- 2 Broad Learning System -- 3 NR IQA Based BLS -- 3.1 Structure Features -- 3.2 Statistical Features -- 4 Numerical Experiments -- 4.1 Single Database Performance Comparisons -- 4.2 Single Distortion Type Performance Comparison -- 4.3 Cross-Validation Performance -- 5 Conclusion -- References -- Hindsight Curriculum Generation Based Multi-Goal Experience Replay -- 1 Introduction -- 2 Related Works -- 3 Background -- 3.1 Multi-Goal RL -- 3.2 Universal Value Function Approximators -- 3.3 Hindsight Experience Replay -- 4 Hindsight Curriculum Generation -- 4.1 Challenges of Hindsight -- 4.2 Clustering and Evaluation -- 4.3 Goals Replay -- 4.4 Multi-Goal Experience Replay -- 5 Experiments -- 6 Conclusion -- References -- Gesture-Based Autonomous Diving Buddy for Underwater Photography -- 1 Introduction -- 2 Related Work -- 3 Application Scenarios and System Description -- 3.1 Applications Scenarios -- 3.2 System Description -- 4 Gesture-Based Control Design -- 4.1 Framework -- 4.2 Theory and Algorithm Design -- 5 Algorithm and System Experiment.5.1 Algorithm Experiment -- 5.2 System Experiment -- 6 Conclusions -- References -- Adaptive Underwater Image Enhancement via Color Channel Compensation Based on Optical Restoration and Fusion -- 1 Introduction -- 1.1 Recent Method -- 1.2 Underwater Image Formation Model -- 1.3 Strength and Weakness -- 2 Restoration and Enhancement Details -- 2.1 Fast Visibility Restoration Applied on Green Channel -- 2.2 Adaptive Color Channel Compensation -- 2.3 White Balance Before Multiscale Fusion -- 3 Experiments -- 3.1 Qualitative Assessment -- 3.2 Quantitative Comparison -- 4 Conclusion -- References -- A Fast Domain Adaptation Network for Image Super-Resolution -- 1 Introduction -- 1.1 A Subsection Sample -- 2 Related Works -- 2.1 Single Image Super Resolution -- 2.2 Domain Adaptation -- 3 The Proposed Method -- 3.1 Overview of Fast Domain Adaptation Network -- 3.2 The Architecture of Fast Domain Adapation Module -- 3.3 Pretrained SR Model -- 4 Experiments -- 4.1 Experimental Details -- 4.2 Experiments on Real-World Dataset -- 4.3 Experiments on Synthetic Dataset -- 5 Conclusion -- References -- Using Conv-LSTM to Refine Features for Lightweight Image Super-Resolution Network -- 1 Introduction -- 2 Methodology -- 2.1 Network Architecture -- 2.2 Deep Feature Extraction -- 2.3 LSTM Feature Refinement -- 3 Experiments and Results -- 3.1 Experiment Settings -- 3.2 Ablation Studies -- 3.3 Comparisons with Other Methods -- 4 Conclusions -- References -- No-Reference Image Quality Assessment for Contrast Distorted Images -- 1 Introduction -- 2 Proposed Method -- 2.1 Image Sequence -- 2.2 Objective Statistical Attributes -- 2.3 Subjective Perception Attributes -- 3 Experimental Results -- 4 Conclusion -- References -- Multi-scale Deformable Deblurring Kernel Prediction for Dynamic Scene Deblurring -- 1 Introduction -- 2 Related Works -- 3 Proposed Methods.3.1 Overview.LNCS sublibrary.SL 6,Image processing, computer vision, pattern recognition, and graphics ;Volume 12890.Computer graphicsCongressesComputer graphics006.6Peng YuxinMiAaPQMiAaPQMiAaPQBOOK9910502616803321Image and Graphics2504630UNINA