Watermarking scheme evaluation - Algorithms need common benchmarks

Watermarking scheme evaluation - Algorithms need common benchmarks
复制标题

DOI:
10.1109/79.879339
复制
发表时间:
2000-09-01
影响因子:
14.9
通讯作者:
Petitcolas, FAP
Petitcolas, FAP
中科院分区:
工程技术1区
文献类型:
--
作者:
Petitcolas, FAP

文献摘要

被引文献

相似文献

在这篇文章中,我们使用了一个对偶的方法来水印的评估问题,通过分裂成两个(独立的)组的评估标准:功能和保证。第一组代表一组可以使用商定的一系列测试进行验证的需求;第二组是一组评估每个功能的级别。这些水平从零或低到非常高。我们正在研究如何评估配置文件可以定义为不同的应用程序,以及如何重要性抽样技术可以用来评估误报率在一个自动化的方式。
In this article we have used a duality approach to the watermarking evaluation problem by splitting the evaluation criteria into two (independent) groups: functionality and assurance. The first group represents a set of requirements that can be verified using agreed series of tests; the second is a set of levels to which each functionality is evaluated. These levels go from zero or low to very high. We are investigating how evaluation profiles can be defined for different applications and how importance sampling techniques could be used to evaluate the false alarm rate in an automated way.