Advances in Visual Computing

Name: Advances in Visual Computing | 16th International Symposium, ISVC 2021, Virtual Event, October 4-6, 2021, Proceedings, Part II
Brand: Springer
Price: 85.59 EUR
Availability: OnlineOnly

16th International Symposium, ISVC 2021, Virtual Event, October 4-6, 2021, Proceedings, Part II

George Bebis Vassilis Athitsos Tong Yan Manfred Lau Frederick Li Conglei Shi Xiaoru Yuan Christos Mousas Gerd Bruder(Herausgeber*in)

Springer (Verlag)

Erschienen am 2. Dezember 2021

XXXVII, 526 Seiten

E-Book

PDF mit Wasserzeichen-DRM

Systemvoraussetzungen

978-3-030-90436-4 (ISBN)

85,59 €inkl. 7% MwSt.

Systemvoraussetzungen

für PDF mit Wasserzeichen-DRM

E-Book Einzellizenz

Als Download verfügbar

Beschreibung

Weitere Details

Weitere Ausgaben

Inhalt

Intro
Preface
Organization
Keynote Talks
Embodied Perception in-the-Wild
Design Tools for Material Appearance
Guidance-Enriched Visual Analytics: Challenges and Opportunities
Learning and Accruing Knowledge over Time Using Modular Architectures
Combining Brain-Computer Interfaces and Virtual Reality: Novel 3D Interactions and Promising Applications
Direct Estimation of Appearance Models for Image Segmentation
Contents - Part II
Contents - Part I
ST: Medical Image Analysis
Video-Based Hand Tracking for Screening Cervical Myelopathy
1 Introduction
2 Related Work
2.1 Development of Medical Treatments for Cervical Myelopathy
2.2 Expansion of 10-Second Grip and Release Test
2.3 Automatic Screening Methods for Various Diseases
3 Method Design
3.1 Recording Grip and Release Test
3.2 Image Processing
3.3 Pre-processing of the Data
3.4 Two-Class Classification
4 Experiments
4.1 Overview
4.2 Validation for Each Finger and Feature Value
4.3 Validation for the Selected Components
5 Discussion
5.1 Consideration of Validation Results
5.2 Comparison with Other Screening Methods
5.3 Limitations and Future Work
6 Conclusion
References
NeoUNet: Towards Accurate Colon Polyp Segmentation and Neoplasm Detection
1 Introduction
2 Related Work
3 Polyp Segmentation and Neoplasm Detection
4 NeoUNet
4.1 Architecture
4.2 Loss Function
5 Experiments and Discussion
5.1 Benchmark Dataset
5.2 Experiment Setup
5.3 Results and Discussion
6 Conclusion
References
Patch-Based Convolutional Neural Networks for TCGA-BRCA Breast Cancer Classification
1 Introduction
2 Related Literature
3 Methodology
3.1 Dataset
3.2 Preprocessing
3.3 Patch Extraction
3.4 Patch Filtering
3.5 Patch Classification
3.6 Patch Aggregation
3.7 Dataset Augmentation
4 Results and Discussion
4.1 Whole Slide Image Classification
4.2 Patch-Level and Slide-Level Classification After Grid Approach
4.3 Patch-Level and Slide-Level Classification After Nuclei-Guided Filtering
4.4 Voting Methods
4.5 Comparison to Related Studies
5 Conclusions
References
CT Perfusion Imaging of the Brain with Machine Learning
1 Introduction
2 Methods
2.1 Overview
2.2 Dataset
2.3 Pre-processing
2.4 Training and Deployment
3 Experiments
4 Discussion
5 Conclusion
References
Analysis of Macular Thickness Deviation Maps for Diagnosis of Glaucoma
1 Introduction
2 Methods
2.1 Macular OCT Imaging
2.2 Macular Deviation Map Processing
2.3 Dataset Overview
2.4 Machine Learning for Feature Evaluation
3 Results and Discussions
4 Conclusions
References
Pattern Recognition
Variational Conditional Dependence Hidden Markov Models for Skeleton-Based Action Recognition
1 Introduction
2 Proposed Model
2.1 Model Definition
2.2 Model Training
2.3 Inference
3 Experimental Evaluation
3.1 Experimental Details
3.2 Parameter Initialization and Hyperparameter Selection
3.3 States, Mixtures and Temporal Dependencies
3.4 Experimental Results
3.5 Missing Values
4 Conclusions
References
The Unreasonable Effectiveness of the Final Batch Normalization Layer
1 Introduction
2 Background
3 Previous Findings and Existing Hypotheses
3.1 Adding a Final Batch Norm Layer Before the Output Layer
3.2 Experimentation on PlantVillage Dataset Subject to Different Configurations
3.3 Derived Hypotheses
4 Implementation Details and Experimental Results
4.1 Removing the Additional BN Layer During Inference (H-1)
4.2 Impact Level Regarding the Imbalance Ratio (H-2) and the Batch Size (H-3)
4.3 Experimentation on the MNIST Dataset: The Impact of Final BN Layer in Basic CNNs and FC Networks [(H-4),(H-5),(H-6)]
5 Discussion
6 Conclusions
References
Video Analysis and Event Recognition
Cross Your Body: A Cognitive Assessment System for Children
1 Introduction
2 Related Work
3 Data Acquisition and Protocol Definition
4 System Definition
4.1 Feature Extraction
4.2 Action Segmentation
4.3 Analysis
4.4 Conclusion
References
Privacy-Aware Anomaly Detection Using Semantic Segmentation
1 Introduction
2 Related Work
2.1 Custom Feature Methods
2.2 Deep Learning Methods
2.3 Object-Centric Methods
2.4 Privacy Preserving Anomaly Detection
3 The Proposed System
3.1 Segmentation Model Comparison
3.2 Quantitative Comparison of Models
3.3 Segmentation Results
3.4 Processing
4 Results and Discussion
4.1 Anomaly Detection Evaluation Metrics
4.2 Global Accuracy
4.3 Anomaly Specific Accuracy
5 Conclusion
References
Learning Self-supervised Audio-Visual Representations for Sound Recommendations
1 Introduction
2 Related Work
3 Datasets
4 Self-supervised Audio-Visual Representation Learning
4.1 Baseline Model
4.2 Attention-Based Model
4.3 Unlabeled Contrastive Learning
5 Performance Evaluation
5.1 Audio-Visual Correlation Performance
5.2 Sound Recommendation Performance
6 Conclusions
References
Poster
Security Automation Through a Multi-processing Real-Time System for the Re-identification of Persons
1 Introduction
2 Background
2.1 Person Detection and Tracking
2.2 Person Re-identification
3 Research Methodology
3.1 Prototype Pipeline
3.2 Datasets
3.3 Person Detection and Tracking
3.4 Vehicle Detection and Assignment
3.5 Clothes Detection
3.6 Clothes Color Detection
3.7 Histogram Processing and Correlation Metrics
3.8 Final Stage - Re-identification Using Multiple Features
3.9 Ethical Considerations
4 Findings and Discussion of Results
4.1 Single Person Comparison from 2 Separate Feeds
4.2 Multiple Persons Comparison from 2 Separate Feeds
5 Conclusion
References
A Method for Transferring Robot Motion Parameters Using Functional Attributes of Parts
1 Introduction
2 Problem Formulation
3 Parts Function
3.1 ``Part Functions'' of Industrial Parts
3.2 Robot Motion Parameters for Assembly
3.3 Basic Idea of Proposed Method Regarding Functional Attributes of Parts
4 Proposed Method for Transferring Robot Motion Parameters
4.1 Module for Functional-Attribute Recognition
4.2 Module for Transferring of Grasping and Action Points
4.3 Transferring Operating Pattern
5 Experiments on Transferring Robot Motion Parameters
5.1 Setup
5.2 Experimental Results
6 Conclusion
References
Uncooperative Satellite 6D Pose Estimation with Relative Depth Information
1 Introduction
2 Related Works
3 Approach
3.1 Image Simulation
3.2 Network Architecture
4 Experiment
5 Conclusion and Future Work
References
Non-homogeneous Haze Removal Through a Multiple Attention Module Architecture
1 Introduction
2 Related Work
3 Proposed Approach
3.1 Architecture
3.2 Loss Functions
3.3 Instance Normalization
4 Experimental Results
4.1 Datasets
4.2 Training Settings
4.3 Comparisons
5 Conclusions
References
Vehicle Detection and Tracking from Surveillance Cameras in Urban Scenes
1 Introduction
2 Background and Related Works
3 Proposed Solution
4 Results and Discussion
5 Conclusion
References
Towards the Creation of Spontaneous Datasets Based on Youtube Reaction Videos
1 Introduction
2 Related Work
3 Proposed Method
3.1 Obtaining Data
3.2 Pre-processing and Frames Extraction
3.3 Faces and Facial Landmarks Extraction
3.4 Most Intense Moment (MIM)
3.5 Facial Emotion Recognition
3.6 Emotion Recognition Analysis
3.7 Dataset Generated
4 Final Considerations
References
Automated Bite-block Detection to Distinguish Colonoscopy from Upper Endoscopy Using Deep Learning
1 Introduction
2 Related Work
3 Methodology
3.1 Image Preprocessing
3.2 Bite-block CNN and Tongue CNN
3.3 Classifications of Bite-block and Tongue Patches
3.4 Detection of Upper Endoscopy Video
4 Experiments and Results
4.1 Datasets
4.2 Training and Testing
4.3 Performance Metrics and Accuracy Comparison
4.4 Effects of Hue and Saturation
5 Conclusion
References
How Does Heterogeneous Label Noise Impact Generalization in Neural Nets?
1 Introduction
2 Related Works
3 Problem Setup
3.1 Multi-class Classification
3.2 Multi-task Learning
3.3 Multi-label Learning
4 Experiments and Datasets
4.1 Multi-class Classification
4.2 Multi-task Learning
4.3 Multi-label Learning
5 Results
5.1 Multi-class Classification
5.2 Multi-task Learning
5.3 Multi-label Learning
6 Discussion
7 Conclusion
References
A Simple Generative Network
1 Introduction
2 SGN Proposed Model
3 Simulation Results
4 Concluding Remarks
References
Hyperspectral Video Super-Resolution Using Beta Process and Bayesian Dictionary Learning
1 Introduction
1.1 Scene Geometry in Hyperspectral Imaging
1.2 Hyperspectral Video Registration
1.3 Contribution
1.4 Paper Organization
2 Related Works
2.1 Super-Resolution
2.2 Hyperspectral Image Analysis
2.3 Hyperspectral Image and Video Super-Resolution
3 Methodology
3.1 Spatial Super-Resolution for a Hyperspectral Video
3.2 Spectral Video Restoration
4 Experiments
4.1 Experimental Settings
4.2 Emperical Results
5 Conclusion
References
Adding Color Information to Spatially-Enhanced, Bag-of-Visual-Words Models
1 Introduction
2 Literature Review
3 Method
3.1 Early vs. Late Fusing
3.2 Quantization Error Considerations
3.3 Choice of Color Descriptor
3.4 Final Design
3.5 Representative Examples of Spatially-Enhanced BOVW Models
4 Experiments
4.1 The Datasets
4.2 Platform
4.3 Experimental Protocol
5 Results
5.1 Boost in Performance
5.2 Comparison with Alternative Color Descriptors
5.3 Comparison with State-of-the-Art
6 Conclusion
References
A Novel Similarity Measure for Retinal Optical Coherence Tomography Images
1 Introduction
2 Related Work
3 Methods
3.1 Structural Similarity Index Measure
3.2 Multi-scale Structural Similarity Index Measure
3.3 Boundary Extraction
4 Experimental Results and Discussions
5 Conclusions
References
Subspace Discrimination Method for Images Using Singular Value Decomposition
1 Introduction
2 Mathematical Preliminaries
3 Numerical Examples
3.1 Discrimination of Still Images
3.2 Motion Separation Using Orthogonal Decomposition
4 Conclusions
References
Cervical Cancer Detection and Classification in Cytology Images Using a Hybrid Approach
1 Introduction
2 Related Work
3 Methodology
3.1 Dataset
3.2 Experimental Pipeline
4 Results
4.1 Lesion Detection
4.2 Lesion Classification
4.3 Complete System
5 Conclusions
References
Ensemble Learning to Perform Instance Segmentation over Synthetic Data
1 Introduction
2 Related Works
3 Proposal of the Ensemble Technique
3.1 Overview of the Proposed Technique
3.2 Ensemble Algorithm for Instance Segmentation: SISE (Simple Instance Segmentation Ensemble)
3.3 Example of SISE Algorithm to Ensemble the Results of Two Instance Segmentation Mask R-CNN Models
4 Experiments and Results
5 Conclusions
References
Improving Efficient Semantic Segmentation Networks by Enhancing Multi-scale Feature Representation via Resolution Path Based Knowledge Distillation and Pixel Shuffle
1 Introduction
2 Related Works
2.1 Overview
2.2 Search and Training in FasterSeg
3 Resolution Path Based Distillation
3.1 Tutor Model with 1/4 Resolution Path
3.2 Feature Affinity Loss
3.3 Teacher Net Guided Student Net Search and Training
4 Experiments
4.1 Dataset and Implementations
4.2 Experimental Results and Analysis
4.3 Analysis of FA Loss
5 Pixel Shuffle Prediction Module
6 Conclusion
References
Towards Stereoscopic Video Deblurring Using Deep Convolutional Networks
1 Introduction
2 Proposed Method
2.1 3D U-Net Architecture
2.2 Modified PAM Architecture
3 Dataset and Experimental Setup
3.1 Dataset
3.2 Experimental Setup
4 Results and Discussions
5 Conclusions and Future Works
References
Color Point Pair Feature Light
1 Introduction
2 3D Descriptors
2.1 Point Pair Feature (PPF) Descriptor
2.2 Color Point Pair Feature (CPPF) Descriptor
3 CPPFL Proposed Descriptor
3.1 Color Point Pair Feature (PPF) Descriptor
3.2 Creating the CPPFL+ Variant
4 Test with Database Captured with a Kinect Device
4.1 Learning Phase Using the Proposed 3D Descriptors
4.2 Recognition Phase Using the Proposed 3D Descriptors
5 Tests with Own Data Obtained with the RealSense Camera
5.1 RealSense Camera 3D Models
5.2 Recognition with Own Data Captured with the RealSense Camera
6 Testing with Lighting Changes
7 Conclusions
References
Semi Automatic Hand Pose Annotation Using a Single Depth Camera
1 Introduction
2 Related Work
3 Creating Training Data Efficiently
3.1 Selecting the Reference Frames
3.2 Initializing the 3D Joint Location in the Reference Frames
3.3 Initializing the 3D Joint Locations in the Remaining Frames
3.4 Global Optimization
4 Evaluations
4.1 Evaluation on Synthetic Data
4.2 Evaluation on Different Number of Reference Frame Selection
5 Conclusion and Future Work
References
FamSearch: Visual Analysis of Genealogical Data
1 Introduction
2 Related Work
3 FamSearch Visualization Tool
3.1 Graphical User Interface
3.2 Visualizations
3.3 Interaction Techniques
4 Application Examples
4.1 Popularity of a Name over Time
4.2 Dips in the Birth Frequency
4.3 French First Names
5 Conclusion and Future Work
References
Hierarchical Sankey Diagram: Design and Evaluation
1 Introduction
2 Design
2.1 Inflow Variants
2.2 Level-of-Detail Variants
3 Results
3.1 Dataset and Web Application
3.2 Visualization Results
4 Evaluation
4.1 Tutorial
4.2 Survey
4.3 User Study Results
5 Conclusions and Future Work
References
Dynamic Antenna Pattern Visualization for Aviation Safety Incorporating Multipath and Situational Awareness
1 Introduction
2 Background
3 Motivation
4 Methodology
5 Examples
5.1 Case 1: Nominal Antenna Patterns
5.2 Case 2: Animated Antenna Pattern
5.3 Case 3: Antenna Array with Bend
5.4 Case 4: Deformed Antenna Pattern by Scatterer
5.5 Case 5: Visualization of Scattered Field only
6 Conclusions
7 Future Work
References
Computer-Assisted Heuristic Evaluation of Data Visualization
1 Introduction
2 Background and Related Work
3 Methodology
3.1 Overview
3.2 Declarative Visualization Libraries
3.3 Selecting and Classifying Heuristic Rules
3.4 Codifying Heuristic Rules
3.5 Extracting Plot Features
3.6 Computer-Assisted Evaluation
4 Case Studies
5 Limitations
6 Conclusion and Future Work
References
Frame Fields for CAD Models
1 Introduction
1.1 Contribution
1.2 Related Work
2 Formalization and Our Approach Overview
3 Compute a Meshable Orthogonal Frame Field
3.1 Problem Characterization
3.2 Prescribe the Target Field Rotation
3.3 Computing an Orthogonal Frame Field
4 Relax the Orthogonality
5 Results and Applications
5.1 Quad Meshing
6 Conclusion
References
Hierarchical Point Distance Fields
1 Introduction
2 Related Work
3 Data Structure
3.1 Error Bounds
3.2 Properties
4 Implementation
4.1 Choosing Example Points
4.2 Computing the Distance
4.3 Multiple Distance Queries
5 Experimental Results
6 Limitations
7 Conclusion
References
Parallel Sphere Packing for Arbitrary Domains
1 Introduction
2 Related Work
2.1 Dynamic Packings
2.2 Geometric Packings
3 Packing Generation
3.1 Parameters Definition
3.2 Layers Generation
3.3 Spheres in Arbitrary Domain
3.4 Parallelization
4 Results
4.1 Cubic Container
4.2 Arbitrary Containers
4.3 Parallelization
5 Conclusions
References
Building 3D Virtual Worlds from Monocular Images of Urban Road Traffic Scenes
1 Introduction
2 Related Work
3 Our Method
4 Experiments, Results and Discussion
5 Conclusions
References
Bodily Expression of Emotions in Animated Agents
1 Introduction
2 Related Work
3 Methods
3.1 Study 1: Affective Body Gestures Identification
3.2 Study 2: Emotion Recognition
4 Results
5 Discussion and Conclusion
References
Fast Approximation of Color Morphology
1 Introduction
2 Basic Definitions
3 Fast Approximation of Color Morphology
3.1 Fourier-Based Approximative Morphology
3.2 Tonal Shift Analysis
4 Experiments
4.1 Comparison of Computational Times
4.2 Visual Quality Comparison
5 Conclusion
References
Augmented Reality Gamification of Intra- and Production Logistics
1 Introduction
2 Related Work
3 Method
3.1 Hypotheses
3.2 Measurements
3.3 Procedure
3.4 Subjects
4 Findings
4.1 Questionnaires
4.2 Trainees' Performance Results
5 Discussion
5.1 Limitations and Future Research
5.2 Conclusion
References
Virtual Training System Based on the Physiological Cycle of the Potato INIAP Suprema
1 Introduction
2 Problem Formulation
3 System Structure
4 Graphic Environment and Virtual Animation
4.1 3D Environment Development
4.2 Virtual Environment Interface and Animation
5 Experimental Results
6 Conclusions
References
Author Index

Systemvoraussetzungen

Als PDF speichern Als Link merken