Thalmann / Musse Crowd Simulation

2. Auflage 2013
ISBN: 978-1-4471-4450-2
Verlag: Springer
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 305 Seiten

Research into the methods and techniques used in simulating crowds has developed extensively within the last few years, particularly in the areas of video games and film. Despite recent impressive results when simulating and rendering thousands of individuals, many challenges still exist in this area.The comparison of simulation with reality, the realistic appearance of virtual humans and their behavior, group structure and their motion, and collision avoidance are just some examples of these challenges. For most of the applications of crowds, it is now a requirement to have real-time simulations - which is an additional challenge, particularly when crowds are very large.Crowd Simulation analyses these challenges in depth and suggests many possible solutions. Daniel Thalmann and Soraia Musse share their experiences and expertise in the application of: ·         Population modeling·         Virtual human animation·         Behavioral models for crowds·         The connection between virtual and real crowds·         Path planning and navigation·         Visual attention models·         Geometric and populated semantic environments·         Crowd renderingThe second edition presents techniques and methods developed since the authors first covered the simulation of crowds in 2007. Crowd Simulation includes in-depth discussions on the techniques of path planning, including a new hybrid approach between navigation graphs and potential-based methods. The importance of gaze attention - individuals appearing conscious of their environment and of others - is introduced, and a free-of-collision method for crowds is also discussed.

Autoren/Hrsg.

Thalmann, Daniel

Musse, Soraia Raupp

Inhaltsverzeichnis

1;Crowd Simulation;3
1.1;Preface;6
1.1.1;Acknowledgements;8
1.2;Contents;9
2;Chapter 1: Introduction;14
2.1;1.1 Requirements and Constraints for Crowd Modeling;15
2.2;1.2 Crowd Simulation Areas;16
2.3;References;17
3;Chapter 2: State-of-the-Art;21
3.1;2.1 Crowd Dynamics;22
3.2;2.2 Sociological Models of Crowds;22
3.3;2.3 Crowd Simulation;23
3.4;2.4 Behavioral Animation of Groups and Crowds;24
3.5;2.5 Crowd Management Training Systems;27
3.6;2.6 Group Behavior in Robotics and Arti?cial Life;28
3.7;2.7 Environment Modeling for Crowds;28
3.7.1;2.7.1 Environment Models;28
3.7.2;2.7.2 Path Planning;29
3.7.3;2.7.3 Collision Avoidance;31
3.8;2.8 Crowd Rendering;32
3.9;2.9 Crowds in Non-real-time Productions;34
3.10;2.10 Crowds in Games;35
3.11;2.11 Crowd Scenario Authoring;36
3.12;References;37
4;Chapter 3: Modeling of Populations;43
4.1;3.1 Introduction;43
4.2;3.2 Creative Methods;44
4.3;3.3 Body Shape Capture;45
4.4;3.4 Interpolated Techniques;46
4.5;3.5 A Model for Generation of Population;48
4.5.1;3.5.1 De?nition of the Initial Data;50
4.5.2;3.5.2 Choice of a Template;50
4.5.3;3.5.3 De?nition of New Somatotypes;50
4.5.4;3.5.4 Calculation of In?uence of Sample Somatotypes;51
4.5.5;3.5.5 Calculation of Mesh Variation;52
4.5.6;3.5.6 Body Parts' Deformation;53
4.5.7;3.5.7 Results and Discussion;55
4.5.7.1;Microscopic Analysis;55
4.5.7.2;Macroscopic Analysis;58
4.6;3.6 Using Computer Vision to Generate Crowds;58
4.6.1;3.6.1 A Model for Generating Crowds Based on Pictures;59
4.6.1.1;Skeleton Initialization;60
4.6.1.2;Image Segmentation;61
4.6.1.3;Learning the Color Model;62
4.6.1.4;Finding the Silhouette;63
4.6.1.5;Silhouette Processing;65
4.6.1.6;3D Pose Identi?cation;66
4.6.1.7;Virtual Human Reconstruction;67
4.6.2;3.6.2 Results;68
4.7;3.7 Crowd Appearance Variety;69
4.7.1;3.7.1 Variety at Three Levels;70
4.7.2;3.7.2 Color Variety;71
4.7.2.1;Principles of the Method;72
4.7.2.2;HSB Color Spaces;73
4.7.2.3;The Need for Better Color Variety;74
4.7.2.4;Segmentation Maps;75
4.7.2.4.1;Principles of Segmentation;75
4.7.2.4.2;Color Variety Storage;77
4.7.3;3.7.3 Accessories;78
4.7.3.1;Simple Accessories;80
4.7.3.2;Complex Accessories;81
4.7.3.3;Loading and Initialization;82
4.7.3.4;Rendering;83
4.7.3.5;Empty Accessories;85
4.7.3.6;Color Variety Storage;87
4.7.3.7;Scalability;87
4.8;3.8 Final Remarks;90
4.9;References;90
5;Chapter 4: Virtual Human Animation;93
5.1;4.1 Introduction;93
5.2;4.2 Related Work in Locomotion Modeling;94
5.2.1;4.2.1 Kinematic Methods;94
5.2.2;4.2.2 Physically Based Methods;95
5.2.3;4.2.3 Motion Interpolation;96
5.2.4;4.2.4 Statistical Models;98
5.3;4.3 Principal Component Analysis;99
5.3.1;4.3.1 Motion Capture Data Process;99
5.3.2;4.3.2 Full-Cycle Model;100
5.3.2.1;Input Data;100
5.3.2.2;Main PCA;100
5.3.3;4.3.3 Motion Extrapolation;102
5.3.3.1;Second PCA Level (Sub-PCA Level 1);103
5.3.3.2;Third PCA Level (Sub-PCA Level 2);104
5.4;4.4 Walking Model;105
5.4.1;4.4.1 Motion Interpolation and Extrapolation;105
5.5;4.5 Motion Retargeting and Timewarping;107
5.6;4.6 Motion Generation;110
5.6.1;4.6.1 Speed Control;110
5.6.2;4.6.2 Type of Locomotion Control;111
5.6.3;4.6.3 Personi?cation Control;111
5.6.4;4.6.4 Motion Transition;112
5.6.5;4.6.5 Results;112
5.7;4.7 Animation Variety;114
5.7.1;4.7.1 Accessory Movements;115
5.8;4.8 Steering;116
5.8.1;4.8.1 The Need for a Fast Trajectory Control;116
5.8.2;4.8.2 The Seek and Funneling Controllers;117
5.9;4.9 Final Remarks;119
5.10;References;119
6;Chapter 5: Behavioral Animation of Crowds;123
6.1;5.1 Introduction;123
6.2;5.2 Related Work;123
6.3;5.3 Crowd Behavioral Models;126
6.3.1;5.3.1 PetroSim's Behavioral Model;126
6.3.1.1;Knowledge;127
6.3.1.2;Status;127
6.3.1.3;Intentions and Decision Process;128
6.3.1.4;Simplifying the FSMs;129
6.3.1.5;An Example of FSM;129
6.3.1.6;Results;131
6.3.2;5.3.2 A Physically Based Behavioral Model;132
6.3.2.1;Interaction with Environment;133
6.3.2.2;Agents' Perception;134
6.3.2.3;Agents' Decision and Action;135
6.3.2.4;Results;137
6.4;5.4 Crowds Navigation;138
6.4.1;5.4.1 Robot Motion Planning;139
6.4.1.1;Discrete Motion Planning;139
6.4.1.2;Exact Motion Planning;140
6.4.1.3;Sampling-Based Methods;140
6.4.1.4;Reactive Methods;141
6.4.1.5;Multiple Robots;141
6.4.2;5.4.2 Crowd Motion Planning;141
6.4.2.1;Models for Safety Applications;142
6.4.2.2;Models for Entertainment Applications;143
6.4.2.3;Models for Virtual Reality Applications;145
6.4.3;5.4.3 A Decomposition Approach for Crowd Navigation;146
6.4.3.1;Objectives;146
6.4.3.2;Navigation Graphs;148
6.4.3.3;Path Planning with Variety;150
6.4.3.4;Scalable Simulation;152
6.4.4;5.4.4 An Hybrid Architecture Based on Regions of Interest (ROI);154
6.5;5.5 A Collision Avoidance Method Based on the Space Colonization Algorithm;157
6.5.1;5.5.1 The Crowd Model: Biocrowds;158
6.5.1.1;Input;159
6.5.1.2;Initialization;159
6.5.1.3;Computation of the Motion Direction;159
6.5.1.4;Computation of the Velocity Vector;161
6.5.1.5;Elimination of Collision Between Finite-Sized Agents;161
6.5.2;5.5.2 Experimental Results;162
6.5.2.1;Impact of the Density of Markers;162
6.5.2.2;The Shape of Trajectories;164
6.5.2.3;Collision Avoidance;164
6.5.2.4;The Stopping Effects;167
6.5.2.5;Interactive Crowd Control;168
6.6;5.6 Gaze Behaviors for Virtual Crowd Characters;169
6.6.1;5.6.1 Simulation of Attentional Behaviors;170
6.6.1.1;Attention Models;170
6.6.2;5.6.2 Gaze Behaviors for Crowds;171
6.6.2.1;Interest Points;171
6.6.3;5.6.3 Automatic Interest Point Detection;172
6.6.4;5.6.4 Motion Adaptation;173
6.6.4.1;Spatial Resolution;174
6.6.4.2;Temporal Resolution;175
6.7;5.7 Final Remarks;176
6.8;References;176
7;Chapter 6: Relating Real Crowds with Virtual Crowds;181
7.1;6.1 Introduction;181
7.2;6.2 Studying the Motion of Real Groups of People;181
7.2.1;6.2.1 Crowd Characteristics;181
7.2.2;6.2.2 Crowd Events;184
7.2.3;6.2.3 Parameters for Simulating Virtual Crowds Using Real Crowd Information;185
7.2.4;6.2.4 Simulating Real Scenes;185
7.2.4.1;First Sequence: People Passing Through a Door;186
7.2.4.2;Second Sequence: People Waiting and Entering the Train;188
7.3;6.3 Sociological Aspects;189
7.4;6.4 Computer Vision for Crowds;191
7.4.1;6.4.1 A Brief Overview on People Tracking;191
7.5;6.5 An Approach for Crowd Simulation Using Computer Vision;193
7.5.1;6.5.1 Using Computer Vision for People Tracking;194
7.5.2;6.5.2 Clustering of Coherent Trajectories;196
7.5.3;6.5.3 Generation of Extrapolated Velocity Fields;197
7.5.4;6.5.4 Simulation Based on Real Data;198
7.5.5;6.5.5 Some Examples;200
7.6;6.6 Final Remarks;202
7.7;References;203
8;Chapter 7: Crowd Rendering;206
8.1;7.1 Introduction;206
8.2;7.2 Virtual Human Representations;207
8.2.1;7.2.1 Human Template;207
8.2.2;7.2.2 Deformable Mesh;207
8.2.3;7.2.3 Rigid Mesh;209
8.2.4;7.2.4 Impostor;209
8.3;7.3 Architecture Pipeline;210
8.3.1;7.3.1 Human Data Structures;212
8.3.2;7.3.2 Pipeline Stages;214
8.4;7.4 Motion Kits;222
8.4.1;7.4.1 Data Structure;222
8.4.2;7.4.2 Architecture;224
8.5;7.5 Database Management;226
8.6;7.6 Shadows;227
8.7;7.7 Crowd Patches;229
8.7.1;7.7.1 Introduction;229
8.7.2;7.7.2 Patches and Patterns;230
8.7.2.1;Patches;230
8.7.2.2;Patterns;231
8.7.3;7.7.3 Creating Patches;231
8.7.3.1;Patterns Assembly;232
8.7.3.2;Static Objects and Endogenous Trajectories;232
8.7.3.3;Exogenous Trajectories: Case of Walking Humans;232
8.7.4;7.7.4 Creating Worlds;233
8.7.4.1;Assembly of Patches;233
8.7.4.2;Patch Templates;234
8.7.5;7.7.5 Applications and Results;235
8.7.5.1;Results;236
8.8;7.8 Final Remarks;237
8.9;References;237
9;Chapter 8: Populated Environments;239
9.1;8.1 Introduction;239
9.2;8.2 Terrain Modeling;240
9.2.1;8.2.1 Plants and Lakes;242
9.2.2;8.2.2 Sky and Clouds;243
9.3;8.3 Generation of Virtual Environments;243
9.4;8.4 A Model for Floor Plans Creation;245
9.4.1;8.4.1 Treemaps and Squari?ed Treemaps;246
9.4.2;8.4.2 The Proposed Model;248
9.4.2.1;Including Connections Among the Rooms;249
9.4.2.2;Including Corridors on the Floor Plans;250
9.4.2.3;3D House Generation;252
9.4.3;8.4.3 Results;253
9.5;8.5 Informed Environment;256
9.5.1;8.5.1 Data Model;259
9.5.2;8.5.2 Topo Mesh;261
9.6;8.6 Building Modeling;263
9.7;8.7 Landing Algorithms;263
9.8;8.8 Ontology-Based Simulation;264
9.8.1;8.8.1 Using Ontology for Crowd Simulation in Normal Life Situations;266
9.8.2;8.8.2 Applying Ontology to VR Environment;267
9.8.3;8.8.3 The Prototype of UEM;267
9.8.4;8.8.4 Simulation Results;270
9.9;8.9 Real-Time Rendering and Visualization;271
9.10;8.10 Implementation Aspects;273
9.11;8.11 Final Remarks;273
9.12;References;274
10;Chapter 9: Applications: Case Studies;277
10.1;9.1 Introduction;277
10.2;9.2 Crowd Simulation for Virtual Heritage;277
10.2.1;9.2.1 Virtual Population of Worshippers Performing Morning Namaz Prayer Inside a Virtual Mosque;278
10.2.1.1;System Design;278
10.2.1.2;Scenario Creation;279
10.2.2;9.2.2 Virtual Roman Audience in the Aphrodisias Odeon;281
10.2.2.1;Crowd Engine Resume;281
10.2.2.2;High-Fidelity Actors;282
10.2.2.3;Scenario Authoring;283
10.2.2.4;Audience Placement;283
10.2.3;9.2.3 Populating Ancient Pompeii with Crowds of Virtual Romans;284
10.2.3.1;Semantics to Behavior;286
10.2.3.2;Long Term vs Short Term Behaviors;287
10.2.3.3;Results;287
10.3;9.3 Immersion in a Crowd;288
10.4;9.4 Crowdbrush;289
10.4.1;9.4.1 Brushes;291
10.4.1.1;Scenario Management;293
10.4.1.2;Scripting;293
10.4.1.3;Results;295
10.5;9.5 Safety Systems;296
10.6;9.6 Olympic Stadium;297
10.7;9.7 Final Remarks;300
10.8;References;300
11;Book Contribution;302
12;Index;304