Watermarking text document based on statistic property of part of speech string
|更新时间:2024-10-14
|
Watermarking text document based on statistic property of part of speech string
Issue 4, Pages: 108-113(2007)
作者机构:
华中科技大学计算机科学与技术学院 2. 武汉工程大学理学院
作者简介:
基金信息:
DOI:
CLC:TP309.7
Published:2007
稿件说明:
移动端阅览
DAI Zu-xu1, HONG Fan1, CUI Guo-hua1, et al. Watermarking text document based on statistic property of part of speech string[J]. 2007, (4): 108-113.
DOI:
DAI Zu-xu1, HONG Fan1, CUI Guo-hua1, et al. Watermarking text document based on statistic property of part of speech string[J]. 2007, (4): 108-113.DOI:
Watermarking text document based on statistic property of part of speech string
摘要
提出了一个将n元熵方程化为至多(n?1)个一元非线性方程求解的算法
证明了算法的正确性
给出了误差估计。利用词性标记串的统计特性设计了一种基于熵的文本数字水印方案
该方案通过改变词性标记串的概率分布使其熵与嵌入的水印一致。由于水印函数是一类实值函数
其值域仅受计算精度限制
可以大幅度提高水印容量。
Abstract
An algorithm resolving nonlinear entropy equation with multivariable was described and proved firstly
then a new scheme where a text document is marked by accommodating probability distribution of certain part of speech strings to entropy constitutes the watermark was designed.Watermark function in the scheme values in a real number set domi-nated by computing precision
it is helpful to advances watermark capacity significantly.