Digital watermarking method, device and system based on Y-Net

Abstract

The invention discloses a digital watermarking method, device and system based on Y-Net, and the method comprises the steps: inputting a carrier image and a watermark image into a Y-Net generation network, and generating a watermark-containing image; inputting the watermark-containing image into a pre-trained compression loss simulation network to obtain a simulation compression watermark-containing image; discriminating the analog compressed watermark-containing image by adopting a pre-trained discrimination network, and obtaining a discrimination result; meanwhile, inputting the analog compressed watermark-containing image into the extraction network to obtain a reconstructed watermark image; and calculating a total loss function of the whole watermark network according to the watermark-containing image, the reconstructed watermark image and a network discrimination result, optimizing the watermark network by taking minimization of the loss function as a target, and considering that the training is ended when the loss is reduced and kept stable, wherein in the training process, the weights of the discrimination network and the compression loss simulation network are fixed. Compared with the existing algorithm, the method has the advantages of high watermark capacity, high concealment, high safety and high robustness.

Description

technical field [0001] The invention belongs to the technical field of digital watermarking, and in particular relates to a Y-Net-based digital watermarking method, device and system. Background technique [0002] The rapid development of Internet technology has enabled more digital works to be disseminated on the Internet. While facilitating communication, the originality of works has also been tested, and copyright protection has become an issue that people attach great importance to. [0003] Digital watermarking technology can provide protection for the originality of the work according to the characteristics of the work itself. In recent years, a large number of digital watermarking models based on deep learning have emerged, which can not only expand the watermark capacity but also effectively improve transmission security. However, there are still some problems in these methods, such as the watermarked image quality and anti-analysis security are not high enough, an...

AI Technical Summary

Problems solved by technology

However, there are still some problems in these methods, such as the watermarked image quality and anti-analysis security are not high enough, and the robustness against attacks such as JPEG compression is not strong.

Images