Multilevel Visual Secret Sharing

Abstract

This paper presents a novel visual secret sharing (VSS) scheme, called the multilevel VSS scheme, for sharing multiple visual secrets in an image among n participants. In the proposed n-level VSS scheme, the content of an image S is distributed to n levels of secrecy, and encoded to n+1 shares that exhibit the following features: (a) each share does not reveal any secret in S, (b) any k (2 k n+1) shares can be used to reveal k 1 levels of secrecy, (c) all secrets in S can be disclosed when all of the n+1 shares are available and (d) the secrets are revealed by visual inspection of the appropriately stacked shares without any computation.

Randomised visual secret sharing scheme for grey‐scale and colour images

“…PVSS scheme presented in [3][4][5][6] addresses the issues of 'all or nothing'. PVSS can be classified into two types based on recovery of an image as: (i) recovering entire image [3][4][5][6] and (ii) recovering image in pieces [7,8]. The first type of PVSS scheme works on recovering entire image progressively with n shares.…”

“…Here entire image will be available after stacking of all shares but unlike 'all or nothing' both domains have some part of the information available regarding the secret image. Block-based schemes proposed by Wang et al [7,8] have many drawbacks, such as: (i) they work on the principle of expansion of share which increases the size of shares resulting in distortion; (ii) the sharing matrix generated by these method causes irregularities in the basis matrix formation hence restricting the number of users; (iii) it is seen that as the number of participants tends to increase the contrast of image decreases; and (iv) the schemes are not meant for grey-scale and colour images.…”

“…Let z k is greater than or equal to ⌈K(k)/2⌉ − 1 then we will consider middle coefficient as our desire location. Once location of embedding is found we use ExtractData function as shown in (7) to extract the data embedded at that location and Restore as shown in (8) to restore back original coefficient. The extraction algorithm is same as that of embedding except we will use Restore and ExtractData function instead of Embed at receiver end.…”

“…The first method [3][4][5][6] considers the secret image as a whole to generate the shares. On the other hand, the second method [7,8] divides the secret image into non-overlapping blocks which contain the information related to the secret image. Superimposition of any two shares reveals the related blocks of the secret image.…”

An EVCS for Color Images with Real Size Image Recovery and Ideal Contrast Using Bit Plane Encoding

“…Two secret images are encoded into two share images and all the secrets can be revealed by keeping one of two shares and shifting the other share several columns and rows in stacking phase. Wang et al [10] proposed scheme which can encode three secrets into four shares, and decoding can be done by superimposing the transparency of shares with different level of contrast. To overcome the angle restriction of Wu and Chen's scheme [11].…”