A Verifiable Lossless Multiple Secret Images Sharing Scheme

Abstract

In the recent years, the researchers have proposed several schemes for sharing multiple secret images using less number of shares through different operations in order to save storage memory and network bandwidth consumption. However, the existing schemes suffer from one or more of the following limitations, viz. no verification of reconstructed secret images, no support for resistance against loss of single share, support for sharing only 2 secret images, incompatibility for sharing single secret image, lossy reconstruction of secret images and no support for sharing all types (binary, grayscale and color) of images. The proposed scheme overcomes these existing limitations, and shares $u-1$ secret images in $v$ ($v\ge 2$) shares out of which at least $u$ ($2 \le u\le v$) shares are required to reconstruct all $u-1$ secret images. The operations are performed in galois field $GF(2^{8})$ during shares generation and secret images reconstruction. The verification of the reconstructed secret images is done by using a random authentication image. The theoretical analysis, comparison and experimental results are provided to indicate the validity, merits and implementability of the proposed scheme.