Image Copy-Move Forgery Detection: New Tools and Techniques

✍ Scribed by Badal Soni, Pradip K. Das

• Publisher: Springer
• Year: 2022
• Tongue: English
• Leaves: 148
• Series: Studies in Computational Intelligence, 1017
• Category: Library

✦ Synopsis

This book presents a detailed study of key points and block-based copy-move forgery detection techniques with a critical discussion about their pros and cons. It also highlights the directions for further development in image forgery detection. The book includes various publicly available standard image copy-move forgery datasets that are experimentally analyzed and presented with complete descriptions. Five different image copy-move forgery detection techniques are implemented to overcome the limitations of existing copy-move forgery detection techniques. The key focus of work is to reduce the computational time without adversely affecting the efficiency of these techniques. In addition, these techniques are also robust to geometric transformation attacks like rotation, scaling, or both.

✦ Table of Contents

Acknowledgements
Contents
About the Authors
Abbreviations
List of Figures
List of Tables
List of Algorithms
1 Introduction
1.1 Introduction
1.2 Classification of Copy-Move Forgery Detection Techniques
1.3 Motivation
1.4 Contributions of Book
1.5 Organization of Book
References
2 Background Study and Analysis
2.1 Keypoint-Based Copy-Move Forgery Detection Techniques
2.1.1 SIFT Key-point-based CMFD Techniques
2.1.2 SURF Key-point-based CMFD Techniques
2.2 Block-Based Copy-Move Forgery Detection Techniques
2.2.1 CMFD Based on Frequency Domain Features
2.2.2 CMFD Based on Dimensionality Reduction Techniques
2.2.3 CMFD Based on Local Binary Patterns
2.2.4 CMFD Based on Texture Features
2.2.5 CMFD Based on Moment Invariant Features
2.2.6 Miscellaneous CMFD Techniques
2.3 CMFD Databases
2.3.1 CASIA
2.3.2 CoMoFoD
2.3.3 MICC-F220, MICC-F2000, MICC-F8 Multi and MICC-F600
2.3.4 Image Manipulation
2.3.5 Coverage
2.3.6 Columbia
2.4 Comparative Investigation of CMFD Techniques
2.5 Summary
References
3 Copy-Move Forgery Detection Using Local Binary Pattern Histogram Fourier Features
3.1 Introduction
3.2 Local Binary Pattern Histogram Fourier Features (LBP-HF)
3.3 Proposed System
3.3.1 Preprocessing
3.3.2 Overlapping Block Partitioning
3.3.3 Lexicographical Sorting
3.3.4 LBP-HF Descriptors Extraction
3.3.5 Block Matching
3.3.6 Forgery Localization
3.4 Experimental Results and Discussion
3.5 Summary
References
4 Blur Invariant Block-Based CMFD System Using FWHT Features
4.1 Introduction
4.2 Proposed Methodology
4.2.1 Overlapping Block Partitioning
4.2.2 Extraction of FWHT Features
4.2.3 Feature Matching and Forgery Localization
4.3 Experimental Results and Discussion
4.4 Summary
References
5 Geometric Transformation Invariant Improved Block-Based Copy-Move Forgery Detection
5.1 Introduction
5.2 Background Knowledge
5.2.1 Speeded-Up Robust Features (SURF)
5.2.2 Maximally Stable Extremal Regions (MSER)
5.3 Proposed Methodology
5.3.1 Pre-processing
5.3.2 SURF Features Detection and Descriptor Extraction
5.3.3 Descriptor Matching
5.3.4 Eight Connected Neighborhood Region
5.3.5 MSER Region Detection and Matching
5.3.6 Outliers Points Removal and Decision of Forgery
5.4 Experimental Results and Discussions
5.5 Summary
References
6 Key-Points Based Enhanced CMFD System Using DBSCAN Clustering Algorithm
6.1 Introduction
6.2 Background Knowledge
6.2.1 Scale Invariant Feature Transform
6.2.2 Density-Based Clustering Algorithm
6.3 Proposed Methodology
6.3.1 Key-Points Detection and Descriptor Extraction
6.3.2 Descriptor Matching
6.3.3 Clustering and Forgery Detection
6.4 Experimental Results and Discussions
6.4.1 Database
6.4.2 Performance Measures
6.4.3 Robustness Test
6.4.4 Analysis Using MICC-F220 Dataset
6.4.5 Analysis Using MICC-F2000 Dataset
6.4.6 Analysis for Multiple Forgeries Detection
6.4.7 Performance Comparison of Proposed System with Existing Methods
6.5 Summary
References
7 Image Copy-Move Forgery Detection Using Deep Convolutional Neural Networks
7.1 Introduction
7.2 Deep Neural-Network-Based Approach
7.3 Problem Description
7.4 Proposed Methodology
7.4.1 Pre-processing
7.4.2 VGG16
7.4.3 Convolution Layer
7.4.4 Region Proposal Network
7.4.5 Non-maximum Suppression
7.4.6 RoIAlign
7.5 Algorithm
7.6 Dataset Descriptions
7.7 Experimental Results and Discussions
References
8 Oriented FAST Rotated BRIEF and Trie-Based Efficient Copy-Move Forgery Detection Algorithm
8.1 Copy-Move Forgery Detection Pipeline
8.2 Existing Methods
8.3 Problem Description
8.4 Data Collection
8.5 Phases of the Algorithm
8.5.1 Noise Reduction
8.5.2 RGB to Gray-Scale Conversion
8.5.3 Key-Point Detection
8.5.4 Key-Point Descriptor
8.5.5 ORB (Oriented FAST Rotated BRIEF)
8.5.6 Matching the Key-Points
8.5.7 Displaying the Matched Key-Points
8.6 Experimental Results and Discussions
8.7 Database
8.8 Performance Measures
8.8.1 True Positive (TP)
8.8.2 False Positive (FP)
8.8.3 True Negative (TN)
8.8.4 False Negative (FN)
8.8.5 True Positive Rate (TPR)
8.8.6 False Positive Rate (FPR)
8.8.7 Precision
8.8.8 Accuracy
8.9 Analysis Among Various Performance Measures
8.9.1 TPR Versus Threshold
8.9.2 Accuracy Versus Threshold
8.9.3 TPR Versus Key-Points
8.9.4 Accuracy Versus Key-Points
8.9.5 Time Versus Threshold
8.9.6 Time Versus Key-Points
8.10 Analysis Using MICC-F220
8.11 Analysis Using MICC-F2000
8.12 Performance Comparison of Proposed System with Existing Methods
8.13 Comparison of Matching Algorithm
References
9 Summing Up
9.1 Findings
9.2 Future Scope