Passive forensics for copy-move image forgery using a method based on DCT and SVD.

Area of Science:

• Digital Image Processing
• Computer Vision
• Forensic Science

Background:

• Increasing use of sophisticated image editing tools necessitates advanced methods for verifying image authenticity.
• Copy-move forgery, a common image tampering technique, poses challenges for detection, particularly in images with large uniform or similar regions.
• Existing detection methods often lack robustness against post-processing operations and precise localization capabilities.

Purpose of the Study:

• To propose a robust and precise method for detecting copy-move image forgery.
• To enhance the accuracy of forgery detection in the presence of common image distortions.
• To accurately locate tampered regions within forged images.

Main Methods:

• The proposed method utilizes Discrete Cosine Transform (DCT) on overlapping image blocks for robust feature extraction.
• Quantization of DCT coefficients and Singular Value Decomposition (SVD) on sub-blocks are applied to reduce feature dimensionality.
• Feature vectors are lexicographically sorted, and duplicated blocks are identified using a shift frequency threshold.

Main Results:

• The method effectively detects multiple instances of copy-move forgery.
• Precise localization of duplicated regions is achieved.
• The technique demonstrates robustness against various post-processing operations, including Gaussian blurring, Additive White Gaussian Noise (AWGN), and JPEG compression.

Conclusions:

• The proposed DCT and SVD-based method offers a robust solution for copy-move image forgery detection.
• It provides accurate localization of tampered areas, outperforming existing techniques in challenging scenarios.
• The method's resilience to common image distortions makes it suitable for real-world forensic applications.