Method for embedding and extracting digital watermarks into/from image

Abstract

The invention discloses a method for embedding and extracting digital watermarks into / from an image. On the basis of the existing digital water mark embedment and extraction based on vector quantization, at first, a stage-II DWT (Discrete Wavelet Transform) is adopted for the construction of a training vector set; when the robustness of the watermarks is guaranteed to some extent, the invisibility of the watermarks is simultaneously taken into account; the contradiction between the robustness and the invisibility of the watermarks is resolved; secondly, embedment positions of the watermarks are selected through the utilization of a visual model; in such a way, a great deal of watermark information is embedded into a DWT domain in a scientific way; and therefore, the contradiction between the great embedment quantity of the watermarks and the invisibility of the watermarks is solved.

Description

technical field [0001] The invention belongs to the technical field of picture management, and more specifically relates to a method for embedding and extracting a digital watermark in a picture. Background technique [0002] With the advent of the WEB2.0 era, the number of pictures on the Internet has increased exponentially, which provides great convenience to network users and also poses two problems to network regulators: (1) Network supervision The author usually supervises the content of the picture through the file name of the picture, but some picture names may be tampered with to escape supervision; (2), some copyrighted pictures may be illegally used by infringers, but network supervision However, those who cannot detect and stop it in the first place. [0003] The above two problems can be solved by using the technology of embedding digital watermark in the picture. By embedding the descriptive information (such as name, content, etc.) and copyright information ...

AI Technical Summary

Problems solved by technology

These contradictions will directly or indirectly affect the effect of digital watermarking

For example, when embedding a watermark by quantization, if the robustness of the watermark is to be enhanced, the quantization noise will increase accordingly, and the invisibility of the watermark will be weakened, which will affect the use of the watermark carrier (picture, audio, etc.); A large amount of watermark embedding will also weaken the invisibility of the watermark, which will affect the use of the watermark carrier; quantitative embedding of the watermark will make the watermark information difficult to resist the impact of amplitude attacks

Images