Improved schemes for visual cryptography (Q5960282)

A visual cryptography scheme (VCS) for a set of \(n\) participants is a technique to encode a secret image into \(n\) shadow images (also called shares or transparencies) where each participant receives one share with the property that certain sets of \(k\) or more of the shares allow the complete visual recovery of the image by stacking the transparencies, while fewer than \(k\) shares yields no information on the image. Such a system is called a \((k,n)\) threshold CVS. It is assumed that each pixel is either black or white and is divided into \(n\) shares, one for each transparency. Each share is a collection of \(b\) black and white subpixels, and a single pixel can be described by an \(n \times b\) Boolean matrix \(S = [s_{ij}]\) where \(s_{ij} = 1\) if the \(j\)th subpixel in the \(i\)th transparency is black. A pixel will be seen as white if sufficiently many subpixels are white and will be seen as black if not too many of them are white. This work analyzes VCS's in which the reconstruction of black pixels is perfect in the sense that all subpixels associated to a black pixel are black. New constructions are given for \((k,n)\) threshold VCS's that improve on the previously best known constructions. A construction is also given for coloured \((2,n)\) threshold VCS's and for coloured \((n,n)\) VCS's, again improving on previously known constructions.