Research On Reversible Data Hiding On JPEG Images With Multi-objective Optimization

With the continuous development of modern science and technology and the significant improvement of the quality of life,the information in daily life increases dramatically.However,at the same time,more and more information security problems have arisen,such as information leakage,loss,and integrity destruction.How to solve this kind of information security problem has become an important research hot spot,and the emergence of reversible data hiding(RDH)technology,as a solution for this kind of problem,can not only protect the invisibility of information,but also completely recover the original cover.Besides,there are many kinds of carriers in RDH,such as image,text,video and so on.And the JPEG image is the most commonly used digital carrier in network transmission,thus researches on RDH based on JPEG image have received widespread attention.Through theoretical analysis of various characteristics of the existing RDH algorithms on JPEG images,and comparing different performances through diverse experiments,this paper concludes the advantages and disadvantages of different schemes from the experimental results,and summarizes some points that can be improved based on current RDH technologies on JPEG image,which includes that the rate-distortion performance of marked images can be better improved,the file size expansion can be more favorably controlled and the rate-distortion performance and file size expansion should be considered at the same time when designing the RDH algorithm.Based on the summary of various existing algorithms and theoretical analysis,this paper proposes a RDH algorithm for JPEG images based on multi-objective optimization,where the two objectives,rate-distortion performance and file size expansion of the marked images,are simultaneously considered into the designing of algorithm.Considering all the above,a multiobjective optimization mathematical model situatable for RDH on JPEG images is proposed,and the corresponding functions are designed for the two optimized objectives to better measure the performances in the marked image.In order to verify the superiority of the proposed RDH on JPEG Images with multiobjective optimization algorithm in the performances of both rate-distortion and file size expansion,this paper gives sufficient theoretical analysis and also combines a large amount of experiments to prove that the algorithm does well in the optimization of the rate-distortion performance and the reduction of file size expansion of the marked images.In addition,the experimental results show that the algorithm proposed in this paper makes up for the shortcomings of the existing algorithms that the rate-distortion performance and file size expansion are not considered at the same time while designing.