一种提高自然语言文本水印容量的算法  

AN ALGORITHM OF IMPROVING THE CAPACITY OF NATURAL LANGUAGE WATERMARKING

作　　者：黄友荣[1] 徐向阳[1] 吴霞[1] 徐晓静[1] 

机构地区：[1]湖南大学计算机与通信学院,湖南长沙410082

出　　处：《计算机应用与软件》2007年第8期180-182,共3页Computer Applications and Software

摘　　要：自然语言文本水印算法通过对文本句子的语法结构或语义结构进行转换来嵌入水印信息。对句子的语法和TMR(Text Meaning Representation)语义结构进行分析,利用句子语法结构的转换不会改变句子TMR语义结构这一性质将语法水印技术和语义水印技术有效结合起来,提出了一种提高自然语言文本水印嵌入容量的算法。该算法的优点是将控制信息和水印信息分离,并根据每个句子本身的特性动态嵌入相应数量的水印信息。实验表明该算法和原有的语法或语义水印算法相比,水印嵌入容量有一定程度的提高。Nature language watermarking algorithm embeds the information by transforming the syntactic or semantic structure of the sentence. After analyzing the syntactic structure and the Text Meaning Representation semantic structure, the syntactic and the semantic watermarking techniques are combined based on the property that transforming syntactic structure of the sentence does not change the TMR semantic structure, and a new algorithm which can improve natural language watermarking capacity is proposed. Separating the control information from watermarking and embedding corresponding quantity watermarking information according to the sentence dynamic itself, the both are the advantages of this algorithm. The experiment result shows that this algorithm can provide better capacity in embedding watermarking information.

关 键 词：自然语言处理 文本水印 TMR 本体论语义学 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]