Research On Reversible Steganography Method Based On Difference Value

As for the encryption of data,we will first come up with cryptography,which is a mature and highly developed discipline with a strong mathematical foundation as its academic support.Steganography provides another way to encrypt data,by hiding secret information in multimedia files such as images,audio,video,etc.,the receiver can extract the secret information hidden in the picture after receiving,so that it can be transmitted through public channels without suspicion,and the goal of "hiding the encryption itself" was achieved in this way.The common metrics are: unrecognability,security,embedded capacity,and embedding efficiency.Reversible data hiding(Reversible data hiding-RDH)is a steganographic method.When the data receiver performs data extraction,it can not only extract secret information,but also recover the carrier of the message.The “difference value” is the difference between adjacent pixels,the calculation method is related to the block mode of the pixel.Many existing RDH methods are implemented based on the differential values.Reversible data hiding is divided into three categories: lossless compression,histogram translation and difference expansion,while histogram translation is divided into: traditional histogram translation,difference histogram translation and prediction error histogram translation.The innovation of this paper is to propose new RDH methods based on difference histogram translation to achieve the purpose of improving the security and steganography efficiency.The main idea of the method is to maintain the statistical characteristics of the dense image as much as possible when the secret information is embedded,especially for the statistical feature of the difference value histogram,to prevent targeted analysis by steganographer.Based on the idea,two RDH methods SFM_A and SFM_B for maintaining the differential histogram translation of image statistical features are proposed.The idea is that the difference value of the pixel after embedding the secret information is still maintained in the range before embedding.In the first method,each pixel to be embedded is embedded with one bit of secret information,and each of the latter's pixels to be embedded is embedded with two bits of secret information.Different from the traditional RDH,the difference value before embedding is needed as the key for secret information extraction and carrier restoration.The operation of this method increases the complexity of secret information extraction,but improves security.With MATLAB as the operating environment,the experimental results show that both the SFM_A and SFM_B methods can significantly maintain the statistical characteristics of the original image,and significantly improve the peak signal-to-noise ratio of the secret image to the original image,that is,less distortion,better security.As for the comparison between the two methods,the peak signal-to-noise ratio of SFM_A is higher,the image quality is better,and SFM_B can increase the steganographic capacity more,and each has its own advantages and disadvantages.