Research On Image-oriented Data Hiding And Restoration Technology

With the advent of the information era,people are paying more and more attention to information security,which makes image security to become an important research direction in the field of information security.Nowadays,smart phones and social platforms are becoming more and more popular,this make it be more and more convenient to share photos taken with others.However,the openness of social platforms and the various image analysis and processing tools increase the risk of image privacy leakage.Thus,image security technology such as image mosaic is needed to realize the security protection of image privacy regions.In the process of protecting information in privacy region,to resist attacks against the protected privacy region,the pixel data expressing the visual information of an image are generally hidden into the non-privacy region.Under this remote protection strategy,the best hiding location of the private data and the maximum size of the privacy region must be solved first for maximizing the protected region while protecting the security of image information.Second,a good coding scheme for data hiding can improve the peak signal to noise ratio(PSNR)of the stego image and provide a big hiding capacity,so designing a PSNR-controllable coding scheme is an important aspect to improve the quality of information protection;in addition,the real purpose of image protection is to allow image accessors to obtain image information corresponding to their privilege levels.Thus,properly restoring the corresponding image information according to the accessors' privilege levels is also an indispensable task for ensuring the information security of the image.To solve the above issues,we propose a serious of novel models and methods:For the maximum protection size and the optimal carrier selection during the remote protection of the image,the upper bound of the protection region and the optimal carrier selection scheme based on the change of roughness(CR)are proposed,the theoretical analysis and experimental results show that,in the image data hiding and restoration based on the reversible mosaic,this scheme can realize the maximum protection size under the premise of satisfying user's requirements for PSNR of the carrier region,and can effectively restore the image mosaic.For the PSNR-controllable encoding and decoding schemes on the privacy data in the carrier region,a new transformation method of data to be hidden is proposed,and it is proved that the encoding schemes can ensure that the difference between the plaintext and ciphertext data is not more than 1,and that the lossless restoration of the private data can be guaranteed.The experimental results demonstrate that the data embedding and extraction based on the scheme not only guarantee the value of PSNR,but also support the variable hiding capacity,thus automatically adapting the information groups with different lengths.To solve the differentiated image restoration based on users'privilege levels,a selective image restoration technology based on user attributes is proposed.By introducing the attribute preference degree and vote attribute based on users' attributes,the attribute-based privilege calculation method is redesigned,and the selective image information restoration with the fast privilege calculation capability is realized.The experimental results show that using the calculation method proposed in this paper,the fast privilege level calculation and accurate partial image restoration are realized.For the progressive image restoration based on users' attributes,an improved CP-ABE system is proposed.By redesigning the encryption scheme and the private key distribution mode,a CP-ABE that has a shorter constant ciphertext and only one bilinear pairings calculation being needed during decryption is obtained.At the same time,based on the atmospheric scattering model,we achieve the progressively lossless restorating imagewith small space-time complexity.The theoretical analysis and experimental results show that the length of ciphertext and the number of bilinear pairings in our CP-ABE algorithms are better than other algorithms with bilinear pairings,and the decryption time and the PSNR of the highest-level restored image are also significantly improved.