计算机科学
数字水印
水印
可扩展性
嵌入
服务器
客户端
查阅表格
稳健性(进化)
扩频
带宽(计算)
计算机安全
计算机网络
人工智能
数据库
操作系统
基因
频道(广播)
生物化学
图像(数学)
化学
作者
M.U. Celik,Aweke Negash Lemma,Stefan Katzenbeisser,M. van der Veen
标识
DOI:10.1109/tifs.2008.926988
摘要
Today, mass-scale electronic content distribution systems embed forensic tracking watermarks primarily at the distribution server. For limiting the bandwidth usage and server complexity and enhancing scalability, it is preferable to embed the watermark at the client. Embedding in these untrusted clients requires secure embedding methods that do not leak unmarked content or the watermarking secrets. In this work, we propose a secure watermark embedding scheme based on lookup tables for spread-spectrum watermarks, which are robust to noise and can be detected without comparison to the original content. We also develop fast detection mechanisms that make the watermark detection feasible for tracking systems with a large number of clients. Our fast detection algorithm improves detection speed of existing methods by six orders of magnitude in a typical system with millions of clients.
科研通智能强力驱动
Strongly Powered by AbleSci AI