Abstract: Text information hiding technology is applied in protection of copyright and integrity of digital media. Under the processing of word segmentation and part-of-speech tagging, a natural language sentence is transformed into a sequence of part of speech symbol, thus the parity of sequence’s inverse order number is used for information hiding. The properties of sequence’s parity under the transformation of symbols’ exchanging, adding and deleting is proved. According to the information hidden, the sequence of part of speech symbol is properly transformed firstly, and the modification of sentence is directed by the transformed sequence, so the feasibility of sentence’s modification is theoretically ensured. Furthermore, aimless modification on natural language sentence can be avoided.

摘要： 文本信息隐藏技术可应用于数字媒体版权和完整性保护．自然语言文本经词性标注处理后变换为词性标记序列，提出了利用序列逆序数奇偶性隐藏信息的算法．证明了逆序数奇偶性在序列符号对换、增加和删除变换下的性质．根据隐藏信息的要求，先对词性标记序列做适当变换，再在变换后的词性标记序列指导下修改自然语言句子，从理论上保证了可行修改的存在性，并能避免直接在自然语言句子层面上做修改的盲目性．