An effective document watermarking technique
Xiao Zhang · Journal of China Institute of Communications · 2003
The character of the binary Chinese and English document image is analyzed in detail, as a result an effective new method is proposed. First, the original document image is cut into several parts, every of which include only one word, then the important pixels of parts is selected out, some of which is selected randomly to embed the watermark bit. The experiment shows the algorithm does not destroy the quality of image, but it can resist several attacks such as signal processing, noise, at the same time, it can be used to detect and locate the slight changes of the multimedia content.