Research On Key Technologies Of Progressive Image Secret Sharing

Along with the wide application and development of internet and multimedia technology, digital images are easily obtained, transmitted and manipulated. Security of digital images has attracted more attention from scientists and engineers. It is important to protect the sensitive information(especially military deployment plans and personal privacy image) from the malicious behavior.Compared with image encryption and information hiding, secret sharing(SS) is losetolerant as well as can be applied in access control and authentication. This thesis focuses on progressive visual cryptographic scheme(PVCS or progressive visual secret sharing,PVSS) with multiple decryptions. Based on the analysis of the basic principles of SS,previous well-known algorithms in progressive SS, SS with multiple decryptions and their advantages and disadvantages are studied. Aiming to solve the problem of single and lossy decryption in traditional progressive SS, in this thesis, based on the given design concepts of separating maximally in parallel and so on, three progressive SS schemes with multiple decryptions are proposed based on stacking, addition and XOR respectively, and PVCS with meaningful shares and multiple decryptions is given.(1) Aiming to solve the problem of traditional PVSS, i.e., no a general k out of n threshold, in this thesis, according to the proposed design concept of dichotomy and rounding, a new threshold PVSS scheme is proposed. The proposed scheme solves the problem of no a general k out of n mechanism with no pixel expansion. Differently from traditional VSS, the proposed threshold PVSS can gain clearer recovered secret image with more shares. Analyses and experiments are conducted to evaluate the security and e?ciency of the proposed scheme. In comparison with previous schemes, the proposed scheme has superior performances and great potential for prominent prevalence in applications.(2) In order to solve the disadvantage of lossy recovery in traditional progressive secret sharing. In this paper, from the exploited design concept of Maximum Likelihood Estimation(MLE), a novel lossless progressive secret sharing(LPSS) model is defined.Then, a general(1, k out of n) LPSS is proposed using addition recovery based on MLE.The proposed model has a large number of valuable features that previous secret sharing schemes fail to have. The proposed LPSS scheme satisfies P(l, k, n) threshold mechanismwhich means that no perceptual information of the secret data will be revealed when the number of shadows is less than l, the degraded perceptual quality will be recovered when the number of shadows is greater than or equal to l and less than k, the more shadows the better recovered perceptual quality will be, and lossless perceptual quality will be recovered when the number of shadows is greater than or equal to k and less than or equal to n.(3) To overcome the drawbacks of low visual quality and lossy recovery for a general k out of n mechanism in traditional VCS with multiple decryptions. In this paper, by the introduced design concept of information maximization, a threshold VCS with the abilities of OR and XOR decryptions is proposed based on random grids(RG). It has the features of both OR-based VCS(OVCS) and XOR-based VCS(XVCS), which can be applied in wider applications. If a copy machine is not available, the secret could be reconstructed by stacking. On the other hand, if a device having XOR operation is available, the visual quality of the recovered secret image is improved as well as the secret can be reconstructed losslessly for(k, n) mechanism. In comparison with previous methods, the proposed scheme has superior performances and features, such as: obtaining progressive visual quality for(k, n) mechanism. Experiments are conducted to evaluate the security and e?ciency of the proposed scheme.(4) For solving the limitations of low visual quality and serious pixel expansion in traditional VCS with meaningful shares, in this paper, due to the presented design concepts of separating maximally in parallel and distributing homogeneously with small halftone error, we propose a halftone visual cryptographic scheme(HVCS) with minimum auxiliary black pixels(ABPs) distributed homogeneously, which is realized via embedding secret image into meaningful shares in the halftoned processing of the cover images by error diffusion. Secret information pixels(SIPs) are fixed in parallel and separated maximally from each other before the halftoned processing. The proposed scheme obtains admirable visual quality and some preferable advantages compared with related meaningful VCSs. When the adopting VCS has the features of both OVCS and XVCS,the proposed scheme will have multiple decryptions.