Collusion Resistant Watermarking Using Convolutional Encoding and Random Spreading
Abstract
This paper presents a Discrete Wavelet Transform based collusion resistant video watermarking to trace colluders involved in unauthorized video distribution. Our scheme uses Tardos-Skoric codes as fingerprints. To reduce the errors on the fingerprinting codes, we propose a joint scheme that combines pseudo-random spreading sequences and convolutional codes. The performance when the fingerprint embedding and the attack are simulated as a binary symmetric channel proves that the proposed scheme performs better in terms of bit error rate and in terms of colluders tracing using binary attacks. Simulations of a darken attack on the watermarked videos show promising results for low to moderate opacities of the fingerprint embedding.
Origin : Files produced by the author(s)