Computer Vision, Imaging and Computer Graphics Theory and Applications: 14th International Joint Conference, VISIGRAPP 2019, Prague, Czech Republic, ... in Computer and Information Science, 1182)

Preface
Organization
Contents
Computer Graphics Theory and Applications
Synthesis and Validation of Virtual Woodcuts Generated with Reaction-Diffusion
	1 Introduction
	2 Literature Review
		2.1 Synthesis of Woodcuts
		2.2 A Brief Review of Reaction-Diffusion
		2.3 Our Previous Work: Reaction-Diffusion Woodcuts
		2.4 Evaluation in NPR
	3 Methodology
		3.1 Input Images and Manual Control
		3.2 Noise Addition
	4 Results
	5 Evaluation
		5.1 Visual Comparison
		5.2 Qualitative Evaluation
	6 Conclusion and Future Work
	References
Synthesising Light Field Volume Visualisations Using Image Warping in Real-Time
	1 Introduction
	2 Background
		2.1 Light Field Representation
		2.2 Light Field View Synthesis
	3 Light Field Synthesis
		3.1 Volume Depth Heuristics
		3.2 Converting Depth to Disparity
		3.3 Disparity Based Image Warping
		3.4 Convolutional Neural Network
	4 Implementation
		4.1 Speed Improvements
		4.2 Datasets
		4.3 Training Procedure
	5 Experiments
		5.1 Generalisation of Deep Learning
		5.2 Example Synthesised Light Fields
		5.3 Time Performance
	6 Conclusion
	References
Motion Capture Analysis and Reconstruction Using Spatial Keyframes
	1 Introduction
	2 Pose Sampling and Reconstruction
		2.1 Keyframe Selection
		2.2 Interpolation
	3 Multidimensional Projection and Unprojection
		3.1 Projection Techniques
		3.2 Two-Step Projection Schemes
		3.3 Unprojection
	4 Proposal and Implementation
		4.1 Pose Distance and Error Estimation
	5 Visualization
		5.1 Pose Inspection
		5.2 Reconstruction Error Visualization
	6 Projection Optimization
		6.1 Optimization for Other Projection Applications
	7 Motion Capture Compression Evaluation
	8 Final Remarks
	References
Human Computer Interaction Theory and Applications
Involving Hearing, Haptics and Kinesthetics into Non-visual Interaction Concepts for an Augmented Remote Tower Environment
	1 Introduction
	2 Theoritical Backgrounds
		2.1 HCI for Remote Control Towers
		2.2 Hearing as a Way to Interact
		2.3 Providing Information Using Touch
		2.4 Considering the Body to Communicate with the Systems
	3 Identification of Relevant ATC Situations
		3.1 Aircraft Location in Poor Visibility Conditions
		3.2 Abnormal Events Location
		3.3 Runway Incursion Awareness
		3.4 Multiple Radio Calls in Case of Multiple RVT
	4 HCI Techniques for Specific ATC Use Cases
		4.1 Using Kinesthetic Interaction and Audification to Provide Interactive Spatial Sound
		4.2 Spatial Alert as Earcon to Provide Location of Abnormal Events
		4.3 Vibrotactive Feedback to Alert from Critical Events
		4.4 Other Vibrotactile Feedback to Distinguish Between Radio Calls
	5 Method
		5.1 Hypothesis
		5.2 Experimental Protocol
		5.3 Participants
		5.4 Scenarios and Experimental Conditions
		5.5 Apparatus
		5.6 Measurements
	6 Experimental Results
		6.1 Performance Analysis
		6.2 Perceived Workload Analysis
		6.3 Post-experiment Interview Analysis
	7 Discussion
	8 Conclusion
	References
Virtual Reality System for Ship Handling Simulations: A Case Study on Nautical Personnel Performance, Observed Behaviour, Sense of Presence and Sickness
	1 Introduction
	2 State of the Art
	3 Materials and Methods
		3.1 Software
		3.2 Hardware
		3.3 Parameters
		3.4 Procedure
		3.5 Participants
	4 Results
		4.1 Analysis of Performances
		4.2 Analysis of Cybersickness
		4.3 Analysis of the Sense of Presence
	5 Conclusions
	References
A Process Reference Model for UX
	1 Introduction
	2 Background
		2.1 Process Reference Model
		2.2 Capability/Maturity Model
		2.3 Capability/Maturity Models in HCI
	3 Contribution
	4 Methodology
		4.1 Comparison of the Related Work
		4.2 Targeted Literature Review
		4.3 Classification of UX Methods
		4.4 Classification of UX Artifacts
	5 Results
		5.1 Primary UX Lifecycle Processes Within a UX Lifecycle
		5.2 UX Base Practices
		5.3 Supporting UX Methods
	6 Application: Past, Present and Future Work
	7 Conclusion
	References
AAT Meets Virtual Reality
	1 Introduction
	2 Psychological Foundations
		2.1 Dual Process Model
		2.2 Cognitive Bias Modification
		2.3 Hierarchical Model of Approach Avoidance Motivation
		2.4 Approach Avoidance Task
	3 Technical Background
		3.1 Virtual Reality
		3.2 Mobile Applications
		3.3 Game Design Elements
	4 Related Work
	5 Requirement Analysis
		5.1 General Requirements
		5.2 VR Requirements
		5.3 App Requirements
	6 VR Demonstrator Version 1
		6.1 Setup
		6.2 Experiments
	7 VR Demonstrator Version 2
		7.1 Improvements
		7.2 Preliminary Results
	8 Mobile App
	9 Conclusions
	References
Information Visualization Theory and Applications
Orthogonal Compaction: Turn-Regularity, Complete Extensions, and Their Common Concept
	1 Introduction
		1.1 Preliminaries
		1.2 Topology-Shape-Metrics Scheme
		1.3 The Compaction Problem
	2 Compaction Methods
		2.1 Turn-Regularity Approach
		2.2 Complete-Extension Approach
	3 Equivalence
		3.1 Auxiliary Graphs
		3.2 Proof of Equivalence
	4 Runtime Analysis
		4.1 Turn-Regularity Formulation  Complete-Extension Formulation
		4.2 Complete-Extension Formulation  Turn-Regularity Formulation
	5 Summary
	References
A Model for the Progressive Visualization of Multidimensional Data Structure
	1 Introduction
	2 Background and Related Work
		2.1 Data State Reference Model
		2.2 Progressive and Incremental Visual Analytics
		2.3 Interactive and Interpretable ML
		2.4 Unsupervised Incremental Techniques
	3 The Model for Progressive Visualization of Structure
	4 Proof-of-Concept Prototype
		4.1 Functionality Overview
	5 Architecture
	6 Evaluation
	7 Discussion
	8 Conclusions
	References
Visualization of Tree-Structured Data Using Web Service Composition
	1 Introduction
	2 Related Work
	3 Requirements of Hierarchy Visualization Techniques
		3.1 Scope of Processing and Hierarchy Visualization Techniques
		3.2 Use Cases of Hierarchy Visualizations
		3.3 Derived Functional Requirements
	4 Concept for Service Orientation and Composition
		4.1 Service Composition and Integration
		4.2 Derived Operational Data Model
		4.3 Derived Capabilities Model
	5 Hierarchy Visualization API
		5.1 Stakeholder
		5.2 API Design Requirements
		5.3 Routes and Mapping
		5.4 Adaptation Process
	6 Evaluation
		6.1 Treemaps as a Service
		6.2 Use Case Evaluation
		6.3 Comparison to HiViSer API
	7 Conclusions
	References
Breaking the Curse of Visual Analytics: Accommodating Virtual Reality in the Visualization Pipeline
	1 Introduction
	2 Background and Related Work
		2.1 Virtual Reality and VREs
		2.2 Immersion and Presence
		2.3 VR for Information Visualization in VA
		2.4 Uncertainty and Validation
		2.5 Positioning of Our Work
	3 The Curse of Visual Data Exploration
	4 VR in the Visualization Pipeline
		4.1 Embedding the Ideal VR in the InfoVis Pipeline
		4.2 Renouncing the Ideal
	5 Breaking the Curse with Virtual Reality
		5.1 Implications for Dimensions
		5.2 Minimizing the Gap Between the Real World and Data World Using VR
	6 Use Case: Crime Scene Investigation
		6.1 Prototype for 4D Crime Scene Investigation
		6.2 Reconstruction of the Real World
		6.3 Projection of Analysis Results
		6.4 Advantages of Deploying Virtual Reality
		6.5 Initial Expert Feedback
	7 Discussion
	8 Conclusion
	References
Designing a Visual Analytics System for Medication Error Screening and Detection
	1 Introduction
	2 Related Work
	3 Requirement Analysis
		3.1 Interviews with Domain Experts
		3.2 FAERS Data Description
		3.3 Motivating Example of Medication Error Detection
		3.4 Workflow of Reports Review by Domain Experts
		3.5 Current Report Analysis Tools at the FDA
		3.6 MEV Design Rationale
	4 MEV Design Process
		4.1 Radial and Bar Charts
		4.2 Node-Link Diagrams
		4.3 Parallel Coordinates
		4.4 Treemaps
	5 The MEV System
		5.1 The MEV Framework
		5.2 MEV Interface Overview
	6 Evaluation
		6.1 Use Case
		6.2 User Study
	7 Discussion
	8 Conclusion
	References
A Layered Approach to Lightweight Toolchaining in Visual Analytics
	1 Introduction
	2 Related Work
	3 A Layered Approach to Lightweight Toolchaining
		3.1 The Principal Ideas Behind the Approach
		3.2 A Model for Layered Lightweight Toolchaining
		3.3 Combining Tools and Flows
		3.4 Interface Ensembles
	4 Providing the Right Tool at the Right Time
		4.1 An Editor for Building a Temporal Toolchain
		4.2 A Temporal Toolchain in Retinal Data Analysis
	5 Providing the Right Data at the Right Time
		5.1 A Library for Data Exchange Between VA Tools
		5.2 A Cross-Tool Analysis Scenario for Astronomical Data
	6 Providing the Right Parameters at the Right Time
		6.1 A Comprehensive Application for the Visualization of Different Presets
		6.2 Toolchain and Presets for the Detection of Cardio-Vascular Anomalies
	7 Conclusion
	References
Fast Approximate Light Field Volume Rendering: Using Volume Data to Improve Light Field Synthesis via Convolutional Neural Networks
	1 Introduction
		1.1 Motivation
		1.2 Problem Statement
		1.3 Contributions
	2 Related Work
		2.1 Neural Network Viewpoint Synthesis
		2.2 Learning from Volumes
		2.3 Rank Pooling
		2.4 Light Field Volume Rendering
	3 Method
		3.1 Viewpoint Synthesis CNN
		3.2 Volumetric Representations
	4 Implementation
	5 Evaluation
		5.1 Qualitative
		5.2 Quantitative
		5.3 Performance Speed
	6 Conclusions
	References
A Reproducibility Study for Visual MRSI Data Analytics
	1 Introduction
	2 Data Description
		2.1 Data Acquisition
		2.2 Quantification of Metabolites
		2.3 Application Scenario
	3 Existing Studies and Tools
		3.1 Existing Studies
		3.2 Analysis Tools
	4 Visual MRSI Data Analytics
		4.1 Image-Space Visualization
		4.2 Feature-Space Visualization
		4.3 Coordinated Interaction in Image and Feature Space
		4.4 Application Scenario, Results, and Discussion
	5 Reproducibility Study
		5.1 Methodology
		5.2 Application Scenario, Results, and Discussion
	6 Conclusions and Future Work
	References
Computer Vision Theory and Applications
A Self-regulating Spatio-Temporal Filter for Volumetric Video Point Clouds
	1 Spatio-Temporal Coherence in Volumetric Video
	2 Previous Work
	3 Proposed System
		3.1 Point Cloud Generation and Upsampling
		3.2 Spatio-Temporal Edge-Aware Scene Flow
		3.3 Scene Flow Point Projection
		3.4 Windowed Hausdorff Filter
		3.5 Dynamic Motion Energy Term
		3.6 Spatio-Temporal Density Term
	4 Results
		4.1 Outdoor Handheld Camera Sequences
		4.2 Indoor Studio Sequences
		4.3 Synthetic Data Sequences
		4.4 Flow Initialization
	5 Limitations and Future Work
	6 Conclusions
	References
Modeling Trajectories for 3D Motion Analysis
	1 Introduction
	2 Related Work
	3 Proposed Method
		3.1 Projecting Motions on Manifolds
		3.2 Combination in Manifolds
	4 Experimental Validation
	5 Conclusion
	References
Quantitative Comparison of Affine Feature Detectors Based on Quadcopter Images
	1 Introduction
	2 Overview of Affine-Covariant Detectors
	3 GT Data Generation
		3.1 Rotation
		3.2 Uniform Motion: Front View
		3.3 Uniform Motion: Bottom View
		3.4 Scaling
	4 Outlier Filtering
		4.1 Homography Estimation for Planes Parallel to the Camera Image
		4.2 Outlier Filtering
	5 Evaluation Method
		5.1 Affine Evaluation
		5.2 Fundamental Matrix Evaluation
	6 Comparison
		6.1 Affine Evaluation
		6.2 Fundamental Matrix Evaluation
	7 Conclusions
	References
An MRF Optimisation Framework for Full 3D Reconstruction of Scenes with Complex Reflectance
	1 Introduction
	2 Related Work
		2.1 Overview of Main 3D Reconstruction Approaches
		2.2 Helmholtz Stereopsis
	3 Framework Overview
	4 Full 3D Modelling via Fusion of 2.5D Reconstructions
	5 Full 3D Modelling via Volumetric Optimisation
	6 Experimental Evaluation
		6.1 Datasets
		6.2 Results with Synthetic Scenes
		6.3 Results with Real Scenes
	7 Conclusions
	8 Future Work
	References
Robustifying Direct VO to Large Baseline Motions
	1 Introduction
	2 State of the Art VO Paradigms
	3 Related Work
	4 Proposed System
		4.1 Notation
		4.2 Feature Assisted Direct Monocular Odometry (FDMO)
	5 FDMO on Every Frame
		5.1 Feature Extraction
		5.2 Pose Estimation
		5.3 Local Mapping
	6 Quantitative Evaluation
		6.1 Computational Cost
		6.2 Datasets
		6.3 Frame Drop Experiment
		6.4 Two Loop Experiment
		6.5 Qualitative Assesment
	7 Discussion and Conclusion
	References
Localization and Grading of Building Roof Damages in High-Resolution Aerial Images
	1 Motivation
	2 Related Work
	3 Preliminaries
		3.1 Superpixel Decomposition
		3.2 Training Data Acquisition
	4 Methods
		4.1 Classification with Random Forest
		4.2 Classification with CNNs
		4.3 Damage Grading
	5 Results and Discussion
		5.1 Damage Detection
		5.2 Amount of Necessary Training Data
	6 Ongoing Work: Footprint Alignment
	7 Conclusion and Outlook
	References
Semantic Image Completion Through an Adversarial Strategy
	1 Introduction
	2 Preliminaries and State-of-the-Art Related Work
		2.1 Generative Adversarial Networks
		2.2 Image Inpainting
		2.3 Face Completion
	3 Proposed Approach
		3.1 Adversarial Based Learning of the Data Distribution
		3.2 Semantic Image Inpainting
		3.3 Conditional Face Completion
	4 Experimental Results
		4.1 Qualitative Assessment
		4.2 Quantitative Analysis and Evaluation Metrics
		4.3 Conditional Face Completion
	5 Conclusions
	References
An Enhanced Louvain Based Image Segmentation Approach Using Color Properties and Histogram of Oriented Gradients
	1 Introduction
	2 Related Work
	3 Description of the Approach
		3.1 From Images to Complex Networks
		3.2 From Complex Networks to Communities
		3.3 Merging Homogeneous Regions Algorithm
	4 Implementation and Performance
		4.1 Features for Similarity
		4.2 Combination of Features Strategies
		4.3 Noise Removal
	5 Experimental Evaluation
		5.1 Datasets
		5.2 Evaluation Metrics
		5.3 Results
	6 Conclusion
	References
Vehicle Activity Recognition Using DCNN
	1 Introduction
	2 Background
		2.1 Trajectory Analysis Techniques
		2.2 Trajectory Representation
		2.3 Deep Neural Networks
	3 Theoretical Formulation
		3.1 Encoding Pair-Wise Vehicle Interactions
		3.2 CNN Based Activity Classification
		3.3 Extending Vehicle Activity Classification with Predictive Trajectory Modelling
	4 Evaluation
		4.1 Datasets and Experimental Set-up
		4.2 Validation of Driver Model
		4.3 Results on Vehicle Activity Recognition
		4.4 Results on Datasets from Different Sources
	5 Conclusion
	References
Quantifying Deformation in Aegean Sealing Practices
	1 Introduction
	2 Related Work
	3 Dataset
		3.1 Pre-processing
	4 Correspondence
		4.1 DAISY
		4.2 BOVW
		4.3 Convolutional Autoencoder
	5 Registration
		5.1 Rigid Pre-alignment
		5.2 Non-rigid Finetuning
	6 Evaluation
		6.1 Visualization of Feature Descriptors
		6.2 Evaluation of Synthetic Deformations
		6.3 Evaluation of Hand-Made Impressions
	7 Results
	8 Summary
	References
Author Index