Reversible Data Hiding Technology Based On Prediction-error Expansion Method

The development of network technology brings us convenience but also challenges,and the issue of information security has become more and more important.When it comes to information security,the first thing that comes to mind is traditional cryptography,whose purpose is to hide information that cannot be deciphered but may still be pirated after decryption.To solve this problem,information hiding technology is more secure than cryptography and has been widely studied and used in the field of information security.Due to the particularity of medical,legal,military,and other fields,the distortion of digital images is very strict.The slight distortion of the image may cause serious accidents.Therefore,this requires the hiding algorithm to embed a large amount of information in the carrier and extract it,and it can still be completely restored to the original carrier.This makes the development of the reversible data hiding technology(Reversible data hiding,RDH)more extensive.Because of the above considerations,this article will study the topic of choosing the direction of reversible data hiding.The main research algorithms of RDH technology are roughly divided into lossless compression method,histogram translation method,and difference expansion method.After extensive research by predecessors,based on the difference expansion algorithm,a reversible data hiding algorithm based on prediction error expansion is proposed.A large number of experiments show that the algorithm based on prediction error expansion has better performance.However,algorithms based on prediction error expansion still have the common conflict problem of reversible data hiding algorithms: "When the embedding capacity is high,the carrier image distortion rate is higher.On the contrary,when the embedding capacity is low,the carrier image distortion rate is relatively low." Therefore,how to improve the relationship between the embedding capacity and the carrier image distortion rate is also an important issue in the study of the direction of reversible information hiding.Based on the above ideas,this paper proposes a new type of RDH algorithm based on Prediction-error-expansion(PEE).The innovations of this method are 1.By embedding the information multiple times into the carrier image,it can effectively prevent the attacker from extracting accurate information when the key is unknown.2.The improvement is improved by multiple layers of embedding.This method improves the embedding capacity and significantly reduces the distortion of the image.This method improves the security of steganography and at the same time improves the relationship between the embedding capacity and the distortion rate of the carrier image.Based on this idea,this paper proposes a new RDH method called the I-PEE method.In the experiment,the original carrier image is partitioned in the embedding process.The data can be embedded hierarchically by partitioning.A closed predictor is used to predict the center pixel value by using the peripheral pixels of the center pixel,and then the secret information is embedded in the prediction error value.In the embedded information,the appropriate range is selected to embed the information according to the prediction error histogram generated by the image.The use of this method increases the difficulty of extracting secret information to a certain extent and therefore ensures the security of the information.This paper uses Matlab as the experimental platform to conduct experiments.The experimental performance results are reflected by the image quality index peak signal-to-noise ratio(PSNR)and structural similarity(SSIM).The proof found that the use of our improved predictor has two particularly effective aspects: First,achieve a large amount of information embedding while significantly improving the peak signal-to-noise ratio,thereby achieving a lower image distortion rate,which has a very good effect.Good performance can better ensure the security of steganographic algorithms;second,the computational complexity is low and there is no large time overhead.