Sharing visual secrets in single image random dot stereograms.

Area of Science:

• Computer Science
• Cryptography
• Image Processing

Background:

• Conventional visual cryptography schemes (VCSs) face transmission risks due to suspicious, noise-like shares.
• Existing methods to hide shares in halftones worsen pixel expansion and image quality.

Purpose of the Study:

• To propose a binocular VCS (BVCS) that utilizes single image random dot stereograms (SIRDS) to conceal shares.
• To develop an encryption algorithm for hiding shared pixels within SIRDS, mitigating transmission risks.

Main Methods:

• A novel (2,n)-BVCS is introduced, integrating shares into SIRDS.
• An encryption algorithm modifies SIRDS random dots based on (2,n)-BVCS construction rules.
• An optimization model is proposed to enhance recovered image contrast while meeting SIRDS visual quality requirements.

Main Results:

• The proposed method generates non-pixel-expansion shares for the BVCS.
• Shares are hidden within SIRDS, maintaining a conventional 2D appearance to reduce suspicion.
• The optimization model successfully balances visual quality and contrast for recovered images.

Conclusions:

• The (2,n)-BVCS effectively reduces transmission risks associated with visual cryptography shares.
• Utilizing SIRDS as cover images offers a secure and visually unobtrusive method for share distribution.
• The proposed optimization model ensures high-quality recovered images within the BVCS framework.