Research On Reversible Data Hiding Algorithm In Digital Images

With the rapid development and popularization of Internet technology,the way of people share and transmit information has gradually changed from traditional media to digital multimedia,such as text,image,audio and video.However,while the Internet brings convenience to people's lives,it also brings serious information security issues,such as privacy leaks and copyright infringement.In order to effectively solve data security problems,reversible data hiding(RDH)technology has received extensive attention and research.As a special data hiding technology,it has the biggest difference from the traditional data hiding,that is,the original cover can be recovered completely without any loss after data extraction.Aiming at the frontier problem of reversible data hiding in digital image,this paper proposes and designs reversible data hiding algorithms under three different application scenarios with the support of human visual redundancy and information coding redundancy.The contributions of this dissertation are listed as follows:1.Reversible Data Hiding in AMBTC Images Based on Compression CodingIn order to solve the problem of insufficient performance of the existing reversible data hiding algorithms in AMBTC images,a novel high-performance reversible data hiding algorithm for AMBTC images is proposed.Firstly,the original spatial domain image is converted into the form of triple by AMBTC coding technology.The correlation between the high and low quantization mean values in the triple is compressed by the MED prediction algorithm to make room for embedding additional data.The ICED strategy proposed in the paper divides the prediction error more detailed,so more reasonable coding is implemented.In addition,according to the probability distribution of the prediction error,the BHC technique is used to allocate a reasonable length of the marker code.More reasonable prediction error coding and variable length coding effectively improve the image coding efficiency.Compared with the existing related work,the proposed algorithm maintains high embedding efficiency and greatly reduces the coding length of the image.Therefore,the proposed algorithm is more practical.2.Reversible Data Hiding in JPEG Images Based on Synchronous ScramblingUnder the new application scenario requirements,this paper proposes a reversible data hiding algorithm in JPEG image based on synchronous scrambling.Different from the previous application requirements,the algorithm does not need to consider the distortion problem caused by the data embedding.Instead,the distortion is utilized to visually scramble the original cover to protect the original cover information.For the characteristics that the most non-zero AC coefficients are-land 1,the paper uses histogram embedding to ensure the embedding capacity performance of the algorithm.In addition,since the DC coefficients of adjacent blocks are highly correlated,and the DC coefficients contain most of the texture information of the original cover,the paper compresses and re-encodes the DC coefficients to make room for secret data embedding and visual scrambling.The experimental results show that compared with the existing related work,the proposed algorithm can provide better embedding capacity performance and scrambling effect for JPEG images with different quality factors.3.Reversible Data Hiding in Encrypted Images Based on Two-MSB PredictionIn order to apply to various services provided by cloud computing and cloud storage technologies,this paper proposes a reversible data hiding algorithm in encrypted images based on two-MSB prediction.Compared with the reversible data hiding in traditional plaintext domain,the encrypted signal has less redundant space and is more difficult to process.The two-MSB prediction strategy proposed by the paper can mark the redundant space in the original plaintext domain and transmit it to the encrypted domain,and reserve space for embedding secret data.The embedding of secret data is achieved by replacing the most significant bit and the next most significant bit of each available pixel.In the best case,the embedding capacity can reach 2 bits/pixel.Compared with the existing related work of reversible data hiding in encrypted domain,the proposed scheme has better embedding capacity performance and direct decryption image quality.In addition,the proposed algorithm also has features such as complete reversibility,separability,and zero-error data extraction.The paper summarizes the related research work of reversible data hiding and forecasts the future research direction and development trend.The work of this thesis promotes the development of reversible data hiding to a certain extent.