Research On High-Capacity Reversible Data Hiding Algorithm In Encrypted Images

Reversible data hiding in images is a technique to embed secret data into a cover image with a reversible manner,so that the cover image can be perfectly recovered after the secret data is extracted.However,in some application scenarios,the image need to be encrypted firstly to protect privacy,then the secret data is hidden in the encrypted image.Therefore,the reversible data hiding in encrypted images has become a hot research topic in recent years.Firstly,the research background,application fields and research status of reversible data hiding in encrypted images consisting of encryption technology and data hiding technology are introduced.At the same time,the basic concepts and typical algorithms of encryption and data hiding are also presented.Then,based on existing problems obtained by simulation and analysis of the existing encrypted-image reversible data hiding algorithms,two important performance indexes,such as the recovery image quality and the embedding capacity,are expected to improve.This thesis analyzes that an encrypted-image reversible data hiding algorithm can not achieve lossless recovery because the type of a pixel was predicted only by four adjacent pixels of it.In order to recover the original image without any error,the pixel-type is predicted by combining the four neighborhoods with pixel itself.And the compressed type-mark matrix,which is used to record the smoothness of the pixels,is embedded into the smooth pixels.In the recovery phase,the original image can be recovered perfectly with the help of the extracted type-mark matrix and its neighbor pixels.The experimental results demonstrate that the proposed method can achieve real reversibility,and improve the embedding capacity and the quality of decrypted image.In the above proposed algorithm,the embedding capacity is related to the chessboard partitioning method.In order to further improve the embedding capacity,a reversible data hiding algorithm based on interpolation framework and pixel classification in encrypted images is designed.Before image encryption,the content owner makes one-quarter of the original pixels as the sampled pixels and the remaining pixels as the non-sampled pixels.Then the non-sampled pixels are predicted as smooth pixels or texture pixels,and a type-mark matrix is needed to record the smoothness of non-sampled pixels.The receiver can obtain the original image after two rounds of recovery with the help of the sampled pixels and the type-mark matrix.Compared with the algorithm in the previous chapter,the designed algorithm allows more pixels to participate in pixel classification,which further improved the embedding capacity.Finally,in order to demonstrate the designed algorithm intuitively,a GUI simulation system is designed to verify the validity of the algorithm.