Digital Forensics and Watermarking: 18th International Workshop, IWDW 2019, Chengdu, China, November 2–4, 2019, Revised Selected Papers (Security and Cryptology)

Preface
Organization
Contents
Deep Learning for Multimedia Security
GAN-Based Steganography with the Concatenation of Multiple Feature Maps
	1 Introduction
	2 Related Work
		2.1 ASDL-GAN Model
		2.2 Yang\'s Model
	3 Proposed Scheme Based on Reconstructed GAN
		3.1 The Overall Introduction of Proposed Scheme
		3.2 Generator Reconstruction
		3.3 Discriminator Design
		3.4 Training of Network
		3.5 Practical Steganographic Scheme
	4 Experimental Results and Discussion
		4.1 Experimental Setup
		4.2 Steganographic Results
		4.3 Testing for Different Concatenations of Feature Maps
		4.4 Comparison with State-of-the-Arts
	5 Conclusion
	References
GAN-TStega: Text Steganography Based on Generative Adversarial Networks
	1 Introduction
	2 Related Works
		2.1 Text Generation-Based Steganography
		2.2 GANs for Text Generation
	3 GAN-TStega Model
		3.1 Generator
		3.2 Discriminator
	4 Update Strategy
	5 Information Hiding and Extraction
	6 Experiments and Results
	7 Conclusion
	References
Optimized CNN with Point-Wise Parametric Rectified Linear Unit for Spatial Image Steganalysis
	1 Introduction
	2 The Proposed Architecture
		2.1 Architecture Overview
		2.2 Point-Wise PReLU Activation Function
		2.3 Feature Fusion
	3 Experiment
		3.1 Setup
		3.2 Results
	4 Conclusion
	References
Light Multiscale Conventional Neural Network for MP3 Steganalysis
	1 Introduction
	2 Related Work
		2.1 Multiscale Convolution Structure
		2.2 Convolution Kernel Factorization
		2.3 Residual Architecture
	3 The Architecture of Proposed Network
		3.1 The Input Data and Preprocessing
		3.2 The Sub-net Block
		3.3 The Classifier Part
	4 Experiment
		4.1 Setup of Experiment
		4.2 The Selection of Sub-net Structure
		4.3 The Result of Detecting MP3 Steganography in Different Domains
		4.4 Comparison with Handcraft Features
		4.5 Comparison with CNN-Based Methods
	5 Conclusion
	References
Improving Audio Steganalysis Using Deep Residual Networks
	1 Introduction
	2 Proposed Method
		2.1 WavSResNet Architecture
		2.2 Convolutional Layers
		2.3 Activation Function
		2.4 Pooling Layers
		2.5 Shortcut Components
	3 Experimental Setup
		3.1 Dataset
		3.2 Training Part
		3.3 Testing Part
		3.4 Evaluation Metric
	4 Experiments
		4.1 Experiments on Three Datasets
		4.2 Comparison with the Variants
		4.3 Comparison with Previous Methods
	5 Conclusion
	References
Cover-Source Mismatch in Deep Spatial Steganalysis
	1 Introduction
	2 Related Works
		2.1 Deep Steganalysis
		2.2 Deep Domain Adaptation
	3 Methodology
		3.1 Analysis of Cover-Source Mismatch in Deep Steganalysis
		3.2 J-Net for Cover-Source Mismatch in Deep Steganalysis
	4 Experiment
		4.1 Experimental Settings
		4.2 Validation of Cover-Source Mismatch
		4.3 Texture Complexity Measurement by A-distance
		4.4 J-Net for Cover-Source Mismatch in Deep Steganalysis
		4.5 Parameter Analysis
		4.6 Bonus Experiments
	5 Conclusion
	References
Ensemble Steganalysis Based on Deep Residual Network
	1 Introduction
	2 The Framework of SRNet
	3 Ensemble Learning
		3.1 Extract Base Models
		3.2 Combination Strategy
	4 Experiments
		4.1 Dataset and Settings
		4.2 Result and Discussion
	5 Conclusion
	References
Digital Forensics and Anti-forensics
Weakly Supervised Adaptation to Re-sizing for Image Manipulation Detection on Small Patches
	1 Introduction
	2 Background and Research Problem
	3 Proposed Approach
		3.1 Classification Pipeline
		3.2 Weakly Supervised Adaptation
	4 Experiments
	5 Conclusion
	References
GRU-SVM Model for Synthetic Speech Detection
	1 Introduction
	2 Proposed Scheme
		2.1 Data Preprocessing
		2.2 The Proposed GRU-SVM Model
	3 Experimental Results and Analysis
		3.1 Data Generation and Environment Construction
		3.2 Results and Discussions
	4 Conclusion
	References
Median Filtering Detection of Small-Size Image Using AlexCaps-Network
	1 Introduction
	2 Methods
		2.1 Related Network
		2.2 The Framework of Proposed Network
		2.3 Image Magnification Method
		2.4 Parameters of Proposed Network
	3 Experimental Results
		3.1 Datasets
		3.2 Experimental Results
	4 Conclusion
	References
Double JPEG Compression Detection Based on Markov Model
	1 Introduction
	2 Theoretical Introduction of Quaternion Discrete Cosine Transform for Color Image
		2.1 Representation of Polar Coordinates of Quaternion
		2.2 Quaternion Discrete Cosine Transform
	3 Proposed Method
		3.1 Difference 2-D Arrays
		3.2 Transition Probability Matrix Derived from Difference 2-D Arrays
		3.3 Refinement Process in Amplitude and Three Angles
	4 Experimental Results
	5 Conclusion
	References
Anti-forensics of Image Sharpening Using Generative Adversarial Network
	1 Introduction
	2 Pix2pix
		2.1 The Network Architecture
		2.2 Generator and Discriminator
	3 Experimental Results
		3.1 Datasets
		3.2 Platform and Settings
		3.3 Results
	4 Conclusion
	References
Digital Watermarking
A New JPEG Image Watermarking Method Exploiting Spatial JND Model
	1 Introduction
	2 Related Work
	3 Proposed Method
	4 Experimental Results
	5 Conclusion
	References
Improved DM-QIM Watermarking Scheme for PDF Document
	1 Introduction
	2 Preliminaries
		2.1 DM-QIM Watermarking
		2.2 PDF File Structure
		2.3 Conventional Method
	3 Proposed Method
		3.1 Expected Degree of Distortion
		3.2 Improved DM-QIM Watermarking
		3.3 Embedding Procedure
		3.4 Extraction Procedure
	4 Experimental Results
		4.1 Distortion
		4.2 Robustness
	5 Conclusion
	References
New Paradigm for Self-embedding Image Watermarking with Poisson Equation
	1 Introduction
	2 Watermark Embedding Procedure
		2.1 Laplacian Operator Processing
		2.2 Reference-Bits Generation and Embedding
	3 Content Recovery Procedure
		3.1 Tampering Detection
		3.2 Restoration of Laplacian Bits
		3.3 Tampered Content Recovery
	4 Experimental Results and Comparisons
	5 Conclusions
	References
Information Hiding
Reversible Data Hiding Based on Partitioning the Prediction Values
	1 Introduction
	2 Related Works
		2.1 Li et al.’s PVO
		2.2 Peng et al.’s IPVO
	3 Proposed Method
		3.1 Partitioning the Prediction Values Based on Rhombus Prediction
		3.2 Adaptive Embedding
		3.3 The Embedding and Extraction Algorithms
	4 Experimental Results
	5 Conclusion
	References
A Novel Reversible Data Hiding with Skin Tone Smoothing Effect for Face Images
	1 Introduction
	2 Proposed Method
		2.1 Reversible Data Hiding
		2.2 Reversible Skin Tone Smoothing
		2.3 Data Extraction and Image Restoration
	3 Experiments
		3.1 Comparison with Traditional RDH
		3.2 Compared with Beautification Algorithm
	4 Conclusion
	References
A Novel Lossless Data Hiding Scheme in Homomorphically Encrypted Images
	1 Introduction
	2 Proposed Scheme
		2.1 Image Encryption
		2.2 Data Hiding
		2.3 Extraction of Additional Data and Reconstruction of Image
	3 Experimental Results
	4 Conclusion
	References
Information Hiding Based on Typing Errors
	1 Introduction
	2 Online Text and Typing Error
	3 Related Work
		3.1 BERT and Masked Language Model
		3.2 Text Steganography with Typing Errors
	4 Steganography Method
		4.1 Embedding
		4.2 Extracting
		4.3 Untouchable Words
	5 Experiment and Example
		5.1 Experiment
		5.2 Result
	6 Conclusion
	References
High-Capacity Reversible Data Hiding in Encrypted Images Based on MSB Prediction
	1 Introduction
	2 Proposed Scheme
		2.1 Generation of Location Map
		2.2 Image Encryption
		2.3 Data Embedding in Encrypted Image
		2.4 Data Extraction and Original Image Recovery
	3 Experimental Results and Analysis
		3.1 Analysis of Scrambling Effect
		3.2 Analysis of Reconstructed Image Quality
		3.3 Analysis of Embedding Capacity
		3.4 Comparison and Discussion
	4 Conclusions and Future Work
	References
Steganography and Steganalysis
Image Steganography Using an Eight-Element Neighborhood Gaussian Markov Random Field Model
	1 Introduction
	2 Preliminaries and Prior Work
		2.1 Notations
		2.2 Model in MG
	3 8-GMRF Model
		3.1 Cover Model
		3.2 Stego Image Model
		3.3 Minimizing the KL Divergence
	4 Experimental Results and Analysis
	5 Conclusions
	References
Broadcasting Steganography in the Blockchain
	1 Introduction
	2 Preliminaries
	3 Proposed Method
		3.1 Data Embedding
		3.2 Data Extraction
		3.3 Example
	4 Performance Evaluation
	5 Conclusion
	References
Designing Non-additive Distortions for JPEG Steganography Based on Blocking Artifacts Reduction
	1 Introduction
	2 Preliminaries
	3 Proposed Method
		3.1 Motivation
		3.2 Selection of Modification Directions
		3.3 Embedding with Blocking Artifact Reduction
	4 Experiments
		4.1 Experimental Setup
		4.2 Determination of Steganographic Settings
		4.3 Comparison to Prior Art
	5 Conclusion and Future Work
	References
Halftone Image Steganography with Distortion Measurement Based on Structural Similarity
	1 Introduction
	2 The Proposed Method
		2.1 Halftone Image Structural Similarity
		2.2 Distortion Measurement for STC
	3 General Framework of Embedding and Extraction
		3.1 Embedding Procedure
		3.2 Extraction Procedure
	4 Experiments and Results
		4.1 Objective Vision Imperceptibility
		4.2 Statistical Security Comparisons
	5 Conclusions
	References
A Motion Vector-Based Steganographic Algorithm for HEVC with MTB Mapping Strategy
	1 Introduction
	2 Proposed Steganographic Strategy
		2.1 Motion Trend Top-List
		2.2 MTB Mapping Strategy
		2.3 Minimize Distortion Strategy
		2.4 Security Strategy
	3 A New Steganography Framework in HEVC
		3.1 HEVC Embedding and Coding Integration Framework
		3.2 Steganographic Embedding and Extracting Process
	4 Experiments and Analysis
		4.1 Experiment Setting
		4.2 Embedded Capacity Comparison
		4.3 Video Bit Changing Rate Comparison
		4.4 Video Quality Performance
	5 Conclusion
	References
New Steganalytic Approach for AMR Steganography Based on Block-Wise of Pulse Position Distribution and Neighboring Joint Density
	1 Introduction
	2 Preliminaries
		2.1 AMR Codec
		2.2 FCB Steganography
	3 Feature Construction
		3.1 Pulses Positions 2-D Array
		3.2 Neighboring Joint Density and Block-Based FCB Feature
	4 Experiments and Results
		4.1 Setup
		4.2 Embedding Effect on Pulses Positions Distributions
		4.3 Results and Discussion
	5 Conclusion
	References
A Novel Feature Selection Model for JPEG Image Steganalysis
	1 Introduction
	2 Proposed Method
		2.1 Feature Selection Step
		2.2 Classification Step
	3 Experiments
		3.1 Experimental Data and Setup
		3.2 Experimental Results and Analysis
		3.3 Parameter Sensitivity Analysis
	4 Conclusion
	References
New Steganalytic Features for Spatial Image Steganography Based on Non-negative Matrix Factorization
	1 Introduction
	2 Fundamentals of Threshold Local Binary Pattern
		2.1 Definition of TLBP
		2.2 TLBP Features for Steganalysis
	3 Obtaining Residual Images with Non-negative Matrix Factorization
		3.1 Definition of Non-negative Matrix Factorization
		3.2 Application of NMF in Obtaining Residuals
		3.3 Non-linear Filtering
		3.4 Residual Quantization and Truncation
	4 Feature Extraction
		4.1 Co-occurrence Matrix Formation
		4.2 Feature Aggregation
		4.3 Non-linear Mapping
	5 Performance Analysis
		5.1 Experimental Settings
		5.2 Performance of Each Feature Subset
		5.3 Comparison with Prior Arts
		5.4 Performance of Cover Source Mismatch
		5.5 Performance of Combined Feature Set
	6 Conclusion
	References
IStego100K: Large-Scale Image Steganalysis Dataset
	1 Introduction
	2 Related Dataset
	3 The Construction of IStego100K
		3.1 Source Image Collection
		3.2 Image Preprocessing
		3.3 Information Embedding
		3.4 Overall Details of IStego100K
	4 Experimental
		4.1 Benchmark Methods and Evaluation Metrics
		4.2 Detection Results of Benchmark Methods
	5 Conclusion
	References
BNS-CNN: A Blind Network Steganalysis Model Based on Convolutional Neural Network in IPv6 Network
	1 Introduction
	2 Related Work
		2.1 Network Steganography
		2.2 Network Steganalysis
	3 The Unified Network Steganalysis Model Based on Convolutional Neural Network (BNS-CNN)
		3.1 Constructing Input Matrix
		3.2 Convolutional Layer
		3.3 Softmax Function and Loss Function
	4 Experiment
		4.1 Data Set and Experimental Setup
		4.2 Experimental Results and Analysis
	5 Summary
	References
Authentication and Security
Behavioral Security in Covert Communication Systems
	1 Introduction
	2 Related Works
		2.1 Steganography and Steganalysis Under the ``Prisoners\' Model\'\'
		2.2 Behavioral Concealment and Behavioral Analysis
	3 The Proposed Framework
	4 Experiments and Analysis
		4.1 Data Collection and Analysis
		4.2 Behavioral Steganography Based on Posting Time
	5 Conclusion
	References
Image Visually Meaningful Cryptography Based on Julia Set Generating and Information Hiding
	1 Introduction
		1.1 The Difference Between VC and IVMC
		1.2 Related Work
	2 Julia Sets Image Generating Based on Improved Newton-Raphson Method
		2.1 Improved Newton-Raphson Method
		2.2 Julia Set Images Generating
	3 Image Visually Meaningful Cryptosystem Based on JSI and Information Hiding
		3.1 Diagram of the Cryptosystem
		3.2 Encryption and Decryption Algorithm Steps
	4 Experiment Results and Safety Analysis
		4.1 Experiment Results of the Algorithm
		4.2 Algorithm Performance Analysis
	5 Conclusion
	References
Multi-attack Reference Hashing Generation for Image Authentication
	1 Introduction
	2 Proposed Method
		2.1 Prior Information Set
		2.2 Problem Setup
		2.3 Multi-attack Reference Hashing
	3 Experiments
		3.1 Experiment Setting
		3.2 Perceptual Robustness
		3.3 Discriminative Capability
		3.4 Authentication Results
		3.5 Conclusions
	References
Cheating Detection in (k, n) Secret Image Sharing Scheme
	1 Introduction
	2 Preliminary
		2.1 Thien-Lin’s Secret Image Sharing Scheme
		2.2 Liu-Sun-Yang’s Scheme
	3 The Proposed Scheme
	4 Experiments and Comparisons
		4.1 An Example for Cheating Detection
		4.2 Share and Recover a Secret Image Utilize the Proposed Scheme
		4.3 Comparison
	5 Conclusion
	References
Author Index