Abstract: Text watermarking protects the copyrights of text works by embedding copyright information(watermark) into a text.This paper proposed a text watermarking technique,in which the watermark is embedded by encoding the subject-verbs of the sentences in a text.A watermark message is converted into a string of the hexadecimal Unicode code.With the help of the language technology platform(LTP) of Harbin Institute of Technology,a series of processes are applied to the text to obtain the subject-verbs in the text.Each of the subject-verbs is encoded with one piece of the Unicode string,which achieves the embedding of the watermark.When extracting the watermark,the subject-verbs are obtained from the detected text and decoded according to the codebook generated in the watermark embedding.The corresponding pieces of the Unicode string are taken out from the codebook and put together in correct order.They are then converted back into the original characters to obtain the embedded watermark message.The proposed algorithm has a good nature of concealment and can resist various watermark attacks.

Key words: Digital watermarking,Text watermarking,Subject-verb encoding,LTP