Related Experiment Videos
Forensic analysis of nonlinear collusion attacks for multimedia fingerprinting
H Vicky Zhao1, Min Wu, Z Jane Wang
1Department of Electrical and Computer Engineering, Institute for Systems Research, University of Maryland, College Park, MD 20742, USA. hzhao@eng.umd.edu
Summary
Digital fingerprinting embeds unique data to trace multimedia content. This study analyzes collusion attacks on Gaussian fingerprints, introducing bounded fingerprints and preprocessing techniques to improve detection and reduce distortion.
Area of Science:
- Computer Science
- Information Security
- Digital Forensics
Background:
- Digital fingerprinting is crucial for tracing multimedia content and preventing unauthorized redistribution.
- Collusion attacks, where multiple fingerprinted copies are combined, pose a significant threat to digital fingerprinting systems.
Purpose of the Study:
- To investigate the effectiveness of average and nonlinear collusion attacks on independent Gaussian fingerprints.
- To analyze the impact of these attacks on perceptual quality and explore methods to mitigate distortion.
- To evaluate detection statistics and propose preprocessing techniques for enhanced fingerprint detection under collusion.
Main Methods:
- Analysis of average and basic nonlinear collusion attacks on independent Gaussian fingerprints.
- Introduction and performance evaluation of bounded Gaussian-like fingerprints under collusion.
- Study of commonly used detection statistics and their performance against collusion attacks.
- Development and testing of a preprocessing technique for extracted fingerprints in collusion scenarios.
Main Results:
- Unbounded Gaussian fingerprints can lead to perceivable distortions in fingerprinted copies and post-attack copies.
- Bounded Gaussian-like fingerprints demonstrate improved performance under collusion attacks.
- The proposed preprocessing technique enhances fingerprint detection performance specifically for collusion scenarios.
Conclusions:
- Collusion attacks significantly impact digital fingerprinting systems, causing perceptual distortion.
- Bounded Gaussian-like fingerprints offer a viable solution to mitigate distortion while maintaining robustness against collusion.
- Preprocessing techniques are effective in improving the detection of fingerprints subjected to collusion attacks, enhancing overall system security.