Robust dense-field based copy-move forgery localization using generic radial harmonic Fourier moment invariants

Area of Science:

• Digital Image Forensics
• Computer Vision
• Multimedia Security

Background:

• Image forensics is crucial for verifying multimedia content authenticity.
• Copy-move forgery, a common image manipulation technique, poses significant challenges.
• Existing feature extraction methods often lack accuracy in representing image content.

Purpose of the Study:

• To propose a novel image representation technique for detecting copy-move forgery.
• To enhance the accuracy and robustness of image forensic analysis.
• To introduce the Generic Radial Harmonic Fourier Moment (GRHFM) for distinctive image representation.

Main Methods:

• Development of the Generic Radial Harmonic Fourier Moment (GRHFM) algorithm.
• Exploitation of transform kernels using complex exponentials and trigonometric functions.
• Application of dense-domain-based matching patterns for forgery detection.

Main Results:

• The proposed GRHFM method achieved superior localization performance compared to state-of-the-art techniques.
• High F1 score (92.5%), precision (95.44%), and recall (91.96%) were reported.
• Demonstrated invariance to various image variations, indicating enhanced robustness.

Conclusions:

• The GRHFM algorithm provides a reliable and distinctive image representation for copy-move forgery detection.
• The method exhibits significant improvements in accuracy and robustness.
• Its reasonable computational cost suggests practical applicability in real-world forensic scenarios.