This book provides a systematic overview of watermarking and steganography methods for triangle meshes related to computer graphics and security. The significance of this research has been well recognized by the growing body of work on watermarking, steganography and steganalysis of 3D meshes. With the evolution of the CAD industry and real-world end-user applications such as virtual reality (VR) and 3D printing, 3D meshes have attracted world-wide attention. Besides, the flexible data structure of 3D geometry provides enough space to embed secret information, making it ideal for applications such as copyright protection and covert communication.
Our goal of the book is to allow readers to systematically understand 3D mesh information hiding technology and its applications as a whole. The book outlines comprehensive techniques, including handcrafted and deep learning-based techniques, digital and physical techniques in the literature and provides standard evaluation metrics for triangle meshes. The up-to-date geometrical deep learning and 3D printing-related algorithms are also covered. Offering a rich blend of ideas and algorithms, the book is up-to-date and self-contained. It is an excellent reference resource for users who are seeking 3D mesh watermarking and steganography algorithms, as well as for graduate students and researchers wanting to grasp the frontiers of triangular mesh processing on data hiding.
Author(s): Hang Zhou, Kejiang Chen, Zehua Ma, Feng Wang, Weiming Zhang
Publisher: Springer
Year: 2023
Language: English
Pages: 196
City: Singapore
Foreword by Nenghai Yu
Foreword by Zhenxing Qian
Preface
Acknowledgments
Contents
About the Authors
Acronyms
1 Introduction
1.1 Introduction of Multimedia Security
1.2 Examples of Watermarking Application
1.2.1 Significance of Digital Watermarking for Multimedia Security
1.2.2 Research Status of Image Watermarking Technology
1.2.3 Research Status of Digital Watermarking Technology Based on Deep Learning
References
2 Basic Concepts
2.1 Watermarking
2.1.1 Digital Watermarking
2.1.1.1 Adaptive/Non-Adaptive Watermarking Systems
2.1.1.2 Blind/Semi-blind/Non-blind Watermarking System
2.1.1.3 Evaluation Metric
2.1.2 Physical Watermarking
2.1.2.1 Physical Distortion on Images
2.1.2.2 Physical Distortion Resilient Image Watermarking
2.1.2.3 3D Physical Distortion and Watermarking
2.2 Steganography
2.2.1 Steganographic Coding
2.2.1.1 Problem Formulation
2.2.1.2 Syndrome Trellis Coding
2.2.2 Steganographic Distortion Design
2.2.2.1 Distortion Definition Principle
2.2.3 Deep Learning Based Steganography
2.2.3.1 ASDL-GAN
2.2.3.2 Training Strategy
2.3 Steganalysis
2.3.1 Handcrafted Feature Based Steganalysis
2.3.1.1 Subtractive Pixel Adjacency Matrix (SPAM)
2.3.1.2 Support Vector Machine (SVM)
2.3.1.3 Spatial Rich Models (SRM)
2.3.1.4 Ensemble Classifier
2.3.2 Deep Learning Based Steganalysis
2.3.2.1 The Transition from Traditional Steganalysis to Deep Learning Steganalysis
2.3.2.2 Characteristics of CNN in the Field of Steganalysis
2.4 Triangle Mesh
2.4.1 Dataset
References
3 3D Mesh Watermarking Techniques
3.1 Early 3D Mesh Watermarking Against Digital Attacks
3.1.1 Spatial Domain-Based 3D Mesh Watermarking
3.1.1.1 LSB-Based Method
3.1.1.2 Similarity-Based Method
3.1.1.3 Statistical Embedding-Based Method
3.1.1.4 Other Methods
3.1.2 Transform Domain-Based 3D Mesh Watermarking
3.1.2.1 Laplacian Transform-Based Methods
3.1.2.2 Wavelet Transform-Based Methods
3.1.2.3 Parameter Transform-Based Methods
3.1.2.4 Other Methods
3.2 Deep Neural Network-Based 3D Mesh Watermarking
3.2.1 Deep Template-Based Watermarking for 3D Morphable Models
3.2.1.1 Preliminary: Chebyshev-Based Spectral Convolution
3.2.1.2 Watermark Embedding and Extracting
3.2.1.3 Adversarial Training
3.2.1.4 Network Training
3.2.2 Deep Watermarking for Topology-Agnostic 3D Models
3.2.2.1 Topology-Agnostic GCN
3.2.2.2 Watermark Embedding Sub-Network
3.2.2.3 Attack Layers
3.2.2.4 Watermark Extracting Sub-Network
3.2.2.5 Loss Function
3.3 3D Mesh Watermarking Techniques Against 3D Print–ScanAttacks
3.3.1 Layering Artifact-Based Watermark
3.3.2 Geodesic Distance-Based Local Watermark
3.3.3 Spectral Subspace-Based Watermark
3.3.4 Surface Norm Distribution-Based Watermark
3.3.5 Shape-Based Watermark
3.3.6 Air Pockets-Based Watermark
3.4 3D Physical Watermarking on 3D Printed Objects
3.4.1 Layer Thickness-Based Watermark
3.4.2 Layer Color-Based Watermark
3.4.3 Slicing Parameters-Based Watermark
3.4.4 Terahertz-Based Watermark
3.4.5 Thermal Image-Based Watermark
3.4.6 Reflectance-Based Watermark
3.4.7 Printing Speed-Based Watermark
3.4.8 Infrared-Based Watermark
References
4 3D Mesh Steganography
4.1 Two-State Steganography
4.1.1 Macro Embedding Procedure Model
4.1.2 Multi-level Embedding Procedure Model
4.1.3 Symmetrical Swap Model
4.1.4 Multi-layer Embedding Model
4.1.5 Static Arithmetic Coding Model
4.1.6 Anti-steganalysis Static Arithmetic Coding Model
4.2 LSB Plane Steganography
4.2.1 Gaussian Curvature Model
4.2.2 Truncated Space Steganography Model
4.2.3 Adaptive Steganography Model
4.2.4 Gaussian Model
4.3 Permutation-Based Steganography
4.3.1 Order Encoding Model
4.3.2 Enhanced Order Encoding Model
4.3.3 Binary Tree Model
4.3.4 Coding Tree Model
4.3.5 Maximum Expected Level Tree Model
4.3.6 One-Ring Neighborhood Model
References
5 3D Mesh Steganalysis
5.1 Universal Steganalysis
5.1.1 Universal Steganalysis Framework
5.1.2 YANG208 Features
5.1.3 YANG40 Features
5.1.4 LFS52 Features
5.1.5 LFS64 Features
5.1.6 LFS76 Features
5.1.7 LFS124 Features
5.1.8 Normal Voting Tensor Model
5.1.9 WFS228 Features
5.1.10 Feature Selection Model
5.2 Specific Steganalysis
5.2.1 PCA Transformation Features
5.2.2 Order Permutation Features
5.3 Cover Source Mismatch Problem
References
6 Future Work and Conclusion
6.1 Open Problems for 3D Mesh Watermarking
6.1.1 Robustness to Causality Problem
6.1.2 Robustness to Representation Conversion
6.1.3 3D Printing-Shooting Resilient Watermark on Curve Surface
6.2 Open Problems for 3D Mesh Steganography
6.2.1 Combining Permutation and LSB Domain
6.2.2 Designing Spatial Steganographic Models
6.2.3 Designing Anti-steganalysis Permutation Steganographic Models
6.2.4 Designing Mesh Batch Steganography Models
6.2.5 Designing 3D Printing Steganography Models
6.3 Open Problems for 3D Mesh Steganalysis
6.3.1 Designing Rich Steganalytic Features
6.3.2 Designing Neural Steganalysis Models
6.3.3 Designing Distance Metric Learning for Permutation Steganalysis
6.3.4 Cover Source Mismatch Problem
6.4 Conclusions
References
