Researches On Visual Secret Sharing

In visual secret sharing (VSS) schemes, a pattern or an image can be divided into many shares that are also called shadows and then held by a group of participants. When some legal participants together bring forth their shares without cheat to combine under certain regulations, the original secret pattern or image can be revealed and is discernable by our visual system. VSS has very wide applications in practice such as ubiquitous computing, the blast-off control of nucleus missile, the safekeeping of secret key, the secrecy of dispensatory, etc. This paper focuses on VSS and pays more attentions on some researches such as non-expansible multi-pixel encoding, ideal contrast VSS scheme with reversing, novel VSS scheme with cover images and authentication, general access structure, etc. By in-depth studies, some valuable fruits have been achieved in this paper.In familiar VSS schemes, there is usually only one pixel to be encoded at each encoding step. In this paper, a new encoding method, called self-adaptive non-expansible multi-pixel encoding (SNME), is presented for VSS. This novel method adopts variable encoding length for each step according to the change of pixel type in the secret image. At each encoding step, the number of pixels written into shares is equal to the number of the consecutive pixels of same type during the scanning. SNME is suitable for many access structure including (n, n) or (k, n) threshold structure as well as general access structure with permission subsets and forbidden subsets, and can be applied to encrypt black-white, gray-scale and chromatic images without pixel expansion. What's more, the decryption doesn't need any computation or cryptography knowledge. Additionally, different scanning modes can be well supported in SNME and an appropriate scanning mode will further promote its advantage of encoding efficiency. For all kinds of images, the experimental results show that this novel method can achieve good quality for decrypted images. High encoding efficiency and low complexity are its main advantages.In VSS schemes, the contrast of reconstructed image is much lost and the effective way to improve is to employ reversing operation to reverse black and white pixels as well as to increase encoding runs. A novel VSS scheme with reversing is presented in this paper. It can achieve a really ideal contrast within only [m/h] encoding runs (where m and h are the number of the total columns and the number of the whole-white columns in the basis matrix to encode white pixels, respectively) and no pixel expansion occurs. It encodes the secret image block by block. A block consists of m pixels, which means that m pixels together join into each encoding step. For the familiar access structures and all kinds of images, the experimental results show that by the proposed scheme, the final reconstructed image is exactly same to the original one. What's more, the experimental results, analyses and comparisons show that the proposed scheme is optimal among those schemes with reversing in encoding runs, system capacity and encoding efficiency, etc. Additionally, this paper gives an improved version for Cimato et al's scheme with reversing and a novel scheme with reversing based on a NPBVSS scheme. To achieve an ideal contrast, the improved scheme can cut down the encoding runs from m to (m-h+1). The new scheme based on NPBVSS scheme promotes the encoding efficiency using pixel-block encoding method and removes pixel expansion. The system expenditure can be cut down to 1/m as that of Yang et al's scheme based on (t, n)-NPBVSS scheme.How to cut down the size of shadow even smaller than the size of original image and hiding the noise-like shadow image into a univocal host image is a new direction in VSS fields. In this paper, a novel VSS scheme with cover images and authentication and a novel method called least significant bit(s) with resizable window (LSBs/RW) are proposed, respectively. The proposed scheme can produce shares with cover image whose size is smaller than that of the original image. When reconstruction, the validity of every share can be verified and the decoded image is exactly same to the original one. The proposed LSBs/RW method usually gives 9 candidate pixels from which a best suited one will replace the responding pixel in cover image. In comparison with usual LSBs methods, this new method has higher security, higher image quality and better flexibility. It usually can achieve a tradeoff between hiding capacity and image quality at different occasions. Additionally, this paper also showed a necessary condition on the least average hiding capacity when LSBs/RW applied to the proposed scheme.The most current researches on secret sharing (SS) mainly focus on the perfect access structures with a limited number of access subsets as well as on the share's bounds under an uniformly distribution and the number of the bits required by a share is used as the measurement of the bounds. So the research results are inevitably limited to some extent. Based on general access structures, some generalized information-theoretic results that are suitable for both perfect and non-perfect access structures with an unlimited number of access subsets are presented in this paper. These results are the general conclusions of many current related works and can be used as the basis for deeper researches. Some potential results can be easily deduced from them and the proof or deduction for many known results can be easier and more concise by them. Additionally, using the information entropy of a share as the measurement of the bounds, some generalized bounds that are suitable for all shares and ones that are suitable only for particular shares are given too. The bounds are also the generalization of many current related results under arbitrary probability distributions. Some of them are tighter than those well-known ones.