Distortionless visual multi-secret sharing based on random grid

摘要

Because visual secret sharing (VSS) has already been well defined, visual multi-secret sharing (VMSS) is more practical for various fields. In a VMSS scheme, N secret messages are encoded into two shares that appear to be noise and these are later stacked by rotating, flipping and turning to visually recognize the secrets. VMSS is generally classified as visual cryptography (VC) based VMSS or random grid (RG) based VMSS. However, VC-based VMSS has the three main drawbacks of pixel expansion, the cost of designing a complex codebook, and shape distortion in the decoded secret image. In order to stack shares at any angle during decoding, secrets are encoded into doughnut-shaped shares. However, the traditional doughnut share methods distort the secret after decoding, because the shape of each share is doughnut. This paper proposes a new RG-based VMSS scheme that defines the shape of each share as a pie-shaped share that does not distort secrets, so that multiple secrets are encoded into two pie shares and then decoded by stacking one pie share on another at different angles of rotation. Experimental results demonstrate the efficiency of the proposed method.