A document multi-watermarking based on document characteristics

A multiple watermarking which is applicable to English-Chinese electronic documents of common format is proposed based on document characteristics. Both robust and fragile watermarking are embedded to ensure copyright protection and content authentication. Feature extraction based on word frequency is built to extract feature from host document. Zero-watermarking is introduced into design of robust watermark to ensure transparency and better watermarking capacity. Fragile watermark is constructed by the SHA-1 hash function to be highly sensitive to tamper. Both robust and fragile watermarks can be extracted blindly. Experimental results show that this algorithm has better transparency and tampering sensitivity, and it is robust against attacks of entities such as cutting, copying, pasting, reversing.