Research On Image Reversible Data Hiding Method Based On Low Neighborhood Correlation

With the development of network,big data and multimedia technology,more and more multimedia information is transmitted through the network.People can download digital information conveniently,which provides great convenience for people's lives.But criminals can also easily modify and publish digital information,which not only damages the legitimate rights and interests of copyright owners,but also raises a series of information security issues related to national security,corporate secrets and personal privacy.Information hiding is an effective technical means for covert communication,copyright protection and content integrity authentication.The reversible information hiding technology proposed in recent years can embed specific information into the carrier and allow legitimate users to recover the original carrier without distortion after extracting the information.Therefore,it has attracted wide attention in fields sensitive to image content such as military,medical,and judicial.The thesis focuses on the image reversible information hiding method with low neighborhood correlation as the main research content,and discusses how to improve the embedding capacity and quality of images with low neighborhood correlation.The main innovations of the thesis are as follows:(1)Aiming at the shortcomings of using only fixed partitioning in the PVO algorithm,a reversible information hiding method that splits the image based on the neighborhood correlation and adaptively selects the embedding position is proposed,which can select the optimal more flexible Embedded method.The low-neighbor correlation area is divided into larger blocks,which improves the peak signal-to-noise ratio and ensures the embedding quality.The high-neighbor correlation area is divided into smaller blocks to obtain more embeddable positions,which makes up for the shortcomings of low-neighbor correlation area and low embedding capacity,thereby improving the algorithm's embedding effect in the entire image.Experiments show that this method can achieve better embedding performance when the embedding amount is small.(2)In the image with obvious texture,the adjacent pixel values are greatly different,and the prediction error obtained by using the sorting algorithm is large.This paper proposes a reversible information hiding method based on texture direction recognition for the problem of poor ranking algorithm embedding effect in texture images.First,determine the main texture direction of the texture image by the gradient direction,and then embed data along the main texture direction.Since the pixel values along the texture direction are relatively similar,selecting related pixels according to the texture direction of the image can achieve a better prediction effect.Experiments show that this method can achieve better embedding performance in texture images,especially single texture images.