Research On Reversible Data Hiding Based On Adaptive Block-partition

With the widespread application of smart devices and social media,data security problems such as multimedia data content being tampered with and a large number of users'sensitive data being leaked have become increasingly prominent.Therefore,data hiding has become an important hotspot of current research.However,in special situations such as medical,military and judicial,even slight distortion of cover image is not allowed.So,reversible data hiding is effectively proposed as an important branch of data hiding,which can recover the original digital content losslessly after correctly extracting the hidden data.Reversible data hiding mostly pursues higher visual quality in given embedding capacity.In order to improve the visual quality of the marked image,high-fidelity idea based on image partition is widely valued.However,traditional image partition is mostly based on fixed-sized blocking,which cannot distinguish image complexity well.Therefore,this thesis proposes to use the texture features of the image to adaptively partition on grayscale and color images,so as to further improve the visual quality.The main work of this dissertation is to study the reversible data hiding algorithm based on adaptive block-partition,the specific research content and innovation are summarized as follows:1.A reversible data hiding algorithm for grayscale image based on adaptive block-partition is proposed.In order to effectively reduce shifting distortion in given embedding capacity,the thesis proposes a reversible data hiding method based on adaptive block-partition.This technique is not to divide an image into equal-sized blocks,but by adaptively dividing the cover image into several image blocks of different sizes to establish multiple histograms,and the image blocks of different sizes are processed successively through multiple sorting and implement adaptive payload allocation according to complexity,then the data is embedded into two sides of prediction error histograms to effectively reduce the shifting distortion.This algorithm not only can effectively distinguish the complexity of the image,improve the visual quality of marked images,but also improves the embedding capacity.2.A color image reversible data hiding algorithm based on adaptive block-partition and channel correlation is proposed.To solve the problem that most of the current reversible data hiding of color images directly apply the grayscale image algorithm to each channel of the color image,and cannot to fully utilize the correlation between the color channels.Therefore,this thesis proposed a color image reversible data hiding based on adaptive block-partition method.In order to make full use of the correlation between the channels,the color image was divided into blocks adaptively,and the pixels of the three channels R,G and B were taken into account.In order to make better use of the smooth region and improve the visual quality of color images,an embedding capacity allocation principle based on channel characteristics was proposed to ensure that more information was embedded in the smooth region.Meanwhile,a B-R-G embedding principle was proposed,which accords with the visual quality of human eyes,and choose the channel which has less impact on the visual quality of human eyes.Experimental results show that the algorithm can effectively improve the visual quality of marked images.