Research On Reversible Data Hiding Algorithm Based On Image Interpolation

In the information age,it is easier and easier to acquire,modify and transmit digital images,resulting in a series of problems such as authenticity and integrity verification of digital images.Information hiding technology has developed rapidly because it can provide effective protection for information.As an important branch of information hiding technology,reversible data hiding(RDH)technology processes the cover data through reversible transformations to embed the secret data.It can not only extract the embedded secret data completely,but also realize the lossless recovery of the original cover data.After more than 20 years of research and development,RDH technology has become a valuable part in the field of information hiding.With digital images as the cover,this paper mainly studies the RDH methods based on interpolation technology(IT).Firstly,the classifications,applications and characteristics of information hiding technology are introduced.Secondly,the basic principle of RDH technology and common performance evaluation indicators are introduced,and five typical RDH methods are introduced and discussed respectively.Finally,based on the problems of the existing RDH methods,such as low data embedding capacity and serious image distortion,two methods are proposed to improve the overall data embedding performance of the IT-based RDH methods.1.An adaptive RDH method based on image interpolationThere is room for improvement of the visual quality of the watermarked images of the existing IT-based RDH methods.To solve this problem,an adaptive RDH method based on image interpolation is proposed.Firstly,the proposed image interpolation method is used to generate the interpolated images,so as to improve the visual quality of the interpolated images.Then,all reference pixels and interpolated pixels in each interpolated image block are used to calculate the two estimates of embedded bits,and the maximum embedded bit of each interpolated pixel is selected adaptively according to the two estimates of embedded bits,so as to improve the data embedding capacity as much as possible.Finally,the least significant bit replacement method is used to embed the secret data into the interpolated pixels to improve the visual quality of the watermarked images.The experimental results show that,the proposed method has higher data embedding capacity and better image visual quality.2.An RDH method for interpolated images based on modulo operation and prediction-error expansionIn order to solve the problem of serious image distortion caused by secondary embedding of existing methods,an RDH method for interpolated images based on modulo operation and prediction-error expansion is proposed.Based on the interpolated images and the partition of pixel blocks,the secret data is embedded in interpolated pixels by the modulo operation,and the secret data is embedded in reference pixels by the prediction-error expansion,so as to make full use of all pixels in each block.The experimental results show that,at the given data embedding capacity,the proposed method is superior to the recent similar methods in terms of visual quality of the watermarked images.