Research On High-capacity Reversible Data Hiding In Encrypted Image Based On Optimized Encoding

Reversible data hiding(RDH)can not only accurately extract the hidden secret information,but also recover the original image without distortion when the secret messages are extracted correctly from the carrier.Accordingly,by combing image encryption and RDH technology,reversible data hiding in encrypted image(RDHEI)can comprehensively protect carrier content,secret information and carrier data,and can be widely used in medicine,military and legal scenes.Currently,existing reversible data hiding schemes in encrypted images are difficult to achieve an effective balance of embedding capacity,security,and visual quality of decrypted images.Considering the aforementioned problems,this thesis proposes new solutions to develop research on reversible data hiding of encrypted images.The main contributions are as follows:(1)Facing the low embedding rate problem,a new reversible data hiding scheme in the encrypted images is proposed based on image bit plane cycling exclusive-OR.Firstly,we successively implement the exclusive-OR processing between the high-level plane and its adjacent low-level bit plane of the image to construct a new pre-processed image.Furthermore,lossless compression is performed over the new pre-processed image according to block rearrangement,which can vacate more embedding rooms.After encrypting the compressed image,additional data is embedded in these vacated rooms.Finally,the receiver can perform separable operations for data extraction and image recovery according to the requirements.Experimental results demonstrate that the bit plane cycling exclusive-OR operation is more suitable for lossless compression so that the embedding capacity is significantly improved.Compared with the existing schemes,proposed scheme has better performance in terms of embedding capacity and visual quality of the decrypted image.(2)Since median edge detector(MED)does not have high prediction accuracy in the texture area of the pixel mutation,we propose a reversible data hiding scheme in encrypted image by combining diamond predictor and parameter binary tree labeling.The image owner first uses the diamond predictor to predict the pixel value and generates a prediction error value bitmap.Furthermore,the image owner encrypts the image content with stream cipher and then determines the type of encrypted pixel.When the data hider receives the encrypted image,he can embed the secret messages into the encrypted and deliver it to the recipient.The proposed algorithm employs parameter binary tree labeling to design the recovery algorithm,and thus achieves a significant improvement in the accuracy of pixel prediction and perfectly recovery.Experimental results demonstrate that the proposed scheme achieves a higher embedding capacity comparing with the existing works.This thesis focuses on the research of high-capacity reversible data hiding technology in encrypted images.The corresponding research results effectively promote the practical application of reversible data hiding technology.