针对目前中文文本数字水印嵌入复杂、鲁棒性不强以及容量不足等问题,结合自然语言处理技术,提出了一种基于汉语词性频率的文本零水印算法.算法通过统计不同词性的频次来确定中频词性,将该词性的词语作为文本特征以构造水印,同时定义了一种基于汉字字符串的相似函数,用以确定阈值和检测水印.实验结果表明,所提算法与同类算法相比更强的鲁棒性.%Aiming at the problems of difficult embedding of Chinese text digital watermarking, weak robustness and lack of watermarking capacity, this paper presented a novel Chinese zero-watermarking algorithm based on the frequency of part-of-speech. Through combining the natural language processing technology, this algorithm counted all kinds of part-of-speech to determine the intermediate frequency of part-of-speech, then found out the words belong with the intermediate frequency of part-of-speech as text features to construct watermark. It also defined a string of similarity function based on Chinese characters to determined the threshold and detect watermark. The experimental results indicate that the proposed algorithm has stronger robustness than before.
展开▼