| With the advent of the era of big data,data that was previously limited to private networks can be connected through the Internet,and the storage and transmission of multimedia data has been greatly facilitated.Users can upload files to the cloud to save local space or share files,but the data stored in the cloud is highly likely to be leaked.Therefore,protecting the security of images has become an urgent issue to be solved.For this reason,researchers have combined encryption and information hiding techniques to protect the image content,and the Reversible Data Hiding in Encrypted Image(RDHEI)has thus become a hot research topic.In this thesis,the compression rate,algorithm security and decrypted image quality are the main objects of study,the bit-plane adaptive chunking compression algorithm based on prediction and the RDHEI based on constructive redundancy are proposed.The main research work of this thesis is as follows.A bit-plane adaptive chunking compression algorithm based on prediction is proposed.By analyzing the effects of using mean prediction and median edge detection in Iterative MSBInversion prediction on the fluctuation values of bit planes,and the effects of the reference pixel selection strategy on the number of predicted bit planes 0/1,an improved Iterative MSBInversion prediction algorithm is proposed.By analyzing the influence of different block sizes on coding length,the strategy for adaptively determining the block size of the bit plane is proposed.The results show that the compressible blocks in the predicted bit plane obtained by the improved prediction algorithm are substantially higher,which is more conducive to compression;the bit-plane adaptive chunking algorithm effectively reduces the coding length.Taking UCID image library as an example,the average compression rate is 0.5901,which is smaller than existing algorithms.In addition,the time complexity of the proposed algorithm is similar to that of the existing algorithms,and it is feasible.The RDHEI based on constructive redundancy is proposed.In the image encryption stage,the encrypted image is generated by constructing redundancy rather than transferring redundancy.First,the original image is compressed into a bit stream based on the bit-plane adaptive chunking compression algorithm,and the original image information is stored in the high bit-plane of the constructed encrypted image through the key;subsequently,the redundancy is constructed in the low bit-plane of the encrypted image,and generate the final encrypted image.In the data embedding stage,some bit planes are selected for data embedding.The results show that the algorithm can achieve full reversibility,and the proposed encryption algorithm approximates the implementation of different keys for different images,which can theoretically resist existing known plaintext attacks.In addition,the embedding rate of the proposed algorithm is similar to existing algorithms.Taking Lena as an example,the maximum embedding rate is up to 0.855 bpp when using double-layer differential expansion embedding.when selecting the appropriate number of bit planes for data embedding,the direct decrypted image can be obtained,and the decrypted image quality is high.Taking the Lena as an example,the decrypted image reaches 57.9514 d B at full embedding.Finally,a Graphical User Interface simulation system is designed to visualize and test the reversible data hiding in encrypted image based on constructive redundancy. |