中文摘要：

基于文本的信息隐藏方法是通过对大量汉语句子进行考察和结构分析,抽出以关键词为核心的改写模板,对具有一定结构特点的语句改写。但是经过同义词替换后可能会破坏句子的语言一致性。针对这一缺点,提出了一种语法检测的算法。首先根据语法库统计出词性的可选搭配对进行词性搭配检测是否合理,再检测词语属性的搭配,最后决定是否对词本身有要求进行检测。在C平台下的数值实验表明该检测算法能够有效的处理文本信息隐藏。

英文摘要：

Text information hiding is based on analyzing and studying the structure of large amount sentences in Chinese, extracting paraphrasing templates based on keywords, which can be used to paraphrase some sentences with special structure. The present natural language information hiding approaches based on Chinese synonymous substitution may destroy the semantic coherence by synonymous substitutions. To deal with this disadvantage, an im- proved syntax check algorithm after sentence change is proposed. First, collocation of words basis on Chinese knowl- edge database is checked, then the match of words＇ attribution and words themselves. Numerical experiment based on C shows the effect of this algorithm.