On the difficulty of aligning VSS random grids

• Recover the secret image without the need to align the shares precisely.
• A deviation-tolerant RG-based VSS method without pixel expansion.
• The proposed scheme has higher misalignment tolerance.

Visual secret sharing (VSS) schemes are classified into two categories—visual cryptography-based (VC) and random grid-based (RG)—with the purpose of encoding a secret into several meaningless shared images that can be superimposed to reveal the secret again. VC-based VSS suffers from the problems of pixel expansion and difficulty in alignment. Although the first problem has been solved by RG-based VSS, the latter remains. This paper proposes a new deviation-tolerant RG-based VSS method without pixel expansion. To demonstrate its feasibility, experiments are conducted to show that the reconstructed secret image is still recognizable visually when one shared image shifts by one or more pixels. Compared with related works, the proposed scheme not only has higher misalignment tolerance but also avoids the significant drawbacks in the related work.