Related Experiment Video
Updated: Jun 15, 2026
08:24
Experimental Assessment of Mouse Sociability Using an Automated Image Processing Approach
Published on: May 15, 2016
Risk-distortion analysis for video collusion attacks: a mouse-and-cat game
Yan Chen1, W Sabrina Lin, K J Ray Liu
1Department of Electrical and Computer Engineering, University of Maryland, College Park, MD 20742, USA. yan@umd.edu
Summary
This study analyzes risk-distortion in video fingerprinting attacks. It reveals how colluder behavior impacts detection, crucial for developing robust copyright protection systems against unauthorized redistribution.
Area of Science:
- Digital forensics
- Information security
- Video processing
Background:
- Copyright protection is vital for digital content, especially video shared online.
- Digital fingerprinting is a common method to prevent unauthorized video redistribution.
- Understanding collusion dynamics is key to designing effective anti-piracy systems.
Purpose of the Study:
- To investigate the relationship between risk (detection probability) and distortion in linear video collusion attacks.
- To formulate and solve the optimal linear collusion attack as an optimization problem.
- To model the attacker-detector interaction as a dynamic game to find optimal strategies.
Main Methods:
- Formulating the optimal linear collusion attack as a constrained optimization problem.
- Deriving the optimal risk-distortion curve by varying risk constraints.
- Applying game theory to model the dynamic interaction between attackers and detectors.
- Experimental verification using real video data.
Main Results:
- The study derives an optimal risk-distortion curve for linear collusion attacks with Gaussian fingerprints.
- It demonstrates that attackers can exploit fixed detector strategies to minimize distortion.
- The min-max strategy is identified as the optimal attack approach when detectors are highly effective.
Conclusions:
- The proposed risk-distortion model provides insights into collusion attack dynamics.
- Understanding these dynamics is essential for developing more resilient digital fingerprinting schemes.
- The findings contribute to enhancing copyright protection for video content in public networks.
