Affine Boolean classification in secret image sharing for progressive quality access control

Secret sharing plays an important role in confidential data protection by splitting the data into few noise-like shares or shadows. In the existing essential threshold based secret image sharing scheme, the decoding process depends on the essential as well as on the non-essential components. This paper proposes an essential threshold based progressive secret image sharing scheme. Affine Boolean functions are used to generate 2i noise-like shares in 'i' levels, out of which at least 's' (2i−1) shares are considered as essential and the rest are non-essential shares. Essential shares suffice to reconstruct/decode the secret image with a certain level of recognizability. The contribution of the non-essential part progressively aids on the quality of the reconstructed image. Robustness in term of decoding reliability of the proposed method is also studied against a set of common operations including random gain change on the share images. Extensive simulation results are also shown to validate the efficacy of the proposed method over a large number of existing works.