Digital image watermarking method based on Cauchy statistical modeling

Abstract

The invention discloses a digital image watermarking method based on Cauchy statistical modeling, wherein the method includes steps of selecting a NSST sub-band with biggest host image energy as an optimal sub-band; modifying its coefficient by multiply embedding method, performing reversed NSST on the sub-band of the modified coefficient and other sub-band, and acquiring the watermarking-containing image; secondly, estimating the rotating angle of the watermarking-containing image at the receiving end through two perpendicular axial lines at the DFT frequency spectrum and amplitude center, correcting and restoring the image; performing the NSST on the corrected watermarking-containing image; selecting a high-entropy block of the sub-band with biggest energy and modeling the coefficient by the Cauchy distribution; estimating the model parameter by the coefficient of the part without watermarking part according to the internal coefficient dependency; using the local maximum potential checking method, and designing a local optimal detector to extract specific watermarking information.

Description

technical field [0001] The invention relates to an image watermarking method based on a statistical model, in particular to a digital image watermarking method based on Cauchy statistical modeling, and belongs to the technical field of digital image copyright protection. Background technique [0002] With the rapid development of multimedia and Internet technology, people have entered the information age, digital multimedia products (such as images, audio, video, etc.) Fast access to multimedia resources. However, the resulting multimedia security issues have also become a hot spot for people to pay attention to. Illegal copying and dissemination of digital multimedia products has become easier and easier, which not only damages the interests of copyright owners, but also creates social trust in multimedia security. crisis. As an important branch in the field of information hiding, digital image watermarking technology has emerged as the times require, and it provides an e...

AI Technical Summary

Problems solved by technology

However, the second type of statistical model watermarking method still has the following deficiencies: First, the transform domains currently used are usually wavelet transform, ridgelet transform, and contourlet transform, and these transforms cannot capture important information and information of two-dimensional images well. Second, when modeling the transform domain coefficients, the selected distribution model was not analyzed in depth, and it was not fully proved whether the established model was optimal for the selected transform domain; third, when estimating the model distribution parameters, the use of Watermarked image signals directly perform parameter estimation, which is greatly affected by the watermark embedding strength; fourth, most existing methods use log likelihood ratios to construct maximum likelihood detectors, and do not try to use other statistical testing strategies to construct new detection methods. To improve the accuracy of watermark detection

Images