An (n, n) Threshold Color QR Code-based Visual Secret Sharing Scheme for Privacy Images

Technical Summary: An (n, n) Threshold Color QR Code-based Visual Secret Sharing Scheme for Privacy Images

Problem Statement
Visual Secret Sharing (VSS) schemes are cryptographic techniques designed to encode secret images into multiple distributed shares, allowing for secure transmission over public channels without exposing the original content. While traditional VSS schemes effectively protect binary, grayscale, and color images, they often suffer from two critical limitations in practical applications:

1. Meaningless Shares: Most existing schemes generate shares that appear as random, unstructured noise. These "meaningless" shares can arouse suspicion among potential attackers, reducing the practical security of the transmission.
2. Lossy Recovery or Complexity: Many schemes capable of handling continuous-tone color images either result in imperfect recovery (residual errors) or rely on complex cryptographic computations. Conversely, schemes that achieve lossless recovery often still produce meaningless shares.

The paper addresses the challenge of designing a VSS scheme that simultaneously ensures lossless reconstruction of continuous-tone color images and generates meaningful, innocuous shares that blend into everyday digital environments to avoid detection.

Methodology
The authors propose a novel (n, n)-threshold Visual Secret Sharing Scheme based on Color QR (CQR) codes, termed CQRVSS. The core methodology leverages the color redundancy inherent in QR codes to encrypt secret images.

• Fundamental Concept: Standard QR codes consist of black and white modules. The scheme exploits the fact that QR code decoders rely on luminance contrast rather than specific hues. By replacing black modules with specific "dark" colors and white modules with specific "light" colors, the authors generate CQR codes that remain decodable by standard scanners while carrying hidden data.
• Color Selection and Luminance Analysis:
  • The scheme utilizes the RGB color model. It defines two sets of colors, $E_0$ (dark) and $E_1$ (light), based on luminance calculations (using the YIQ model).
  • To ensure security and prevent partial recovery with fewer than $n$ shares, the authors introduce random values into the color channels. This modifies the base colors (e.g., Red, Green, Blue, Cyan, Magenta, Yellow) to create modified sets $\hat{E}_0$ and $\hat{E}_1$. This ensures that any single share or subset of shares does not reveal information about the secret.
• Encryption Process (Secret Sharing):
  • The scheme requires $n$ cover QR codes (where $n \ge 4$).
  • For each pixel in the secret image, the algorithm selects four distinct cover QR codes.
  • Depending on the combination of black (K) and white (W) modules in the selected cover QR codes at a specific coordinate, the algorithm modifies the corresponding color values in the four shares.
  • The modification involves bitwise XOR operations on the Red, Green, and Blue channels. The process ensures that the XOR sum of all $n$ shares at any pixel position reconstructs the original secret pixel value exactly.
  • The algorithm handles various scenarios (e.g., 4K, 4W, 3K1W, 2K2W, 1K3W) by selecting specific color combinations from the modified sets to maintain high contrast for decoder readability while embedding the secret.
• Decryption Process:
  • The receiver must possess all $n$ shares.
  • Decryption is performed by applying a bitwise XOR operation across all $n$ shares pixel-by-pixel.
  • The result is the lossless recovery of the original continuous-tone color image.

Key Contributions

1. Meaningful Shares: The scheme uses Color QR codes as shares. Unlike random noise, these shares are valid, decodable QR codes that carry public information. This significantly reduces the suspicion of attackers monitoring public channels, enhancing the scheme's practical security.
2. Lossless Recovery: The proposed method achieves perfect, error-free reconstruction of continuous-tone color images. This is critical for applications requiring high fidelity, such as medical imaging, scientific microscopy, and design drawings.
3. (n, n)-Threshold Security: The scheme is designed such that fewer than $n$ shares reveal absolutely no information about the secret. The introduction of random values in the color channels ensures that partial reconstruction attempts yield random, unintelligible results.
4. Simplicity: The encryption and decryption processes rely on simple XOR operations and color substitution, avoiding complex cryptographic computations.

Experimental Results
The authors conducted experiments using MATLAB R2016b with standard test images (Baboon, Barbara) and the ZXing library for QR code generation.

• Recovery Quality: In both (4, 4) and (7, 7) threshold configurations, the reconstructed images were identical to the original secret images. The Peak Signal-to-Noise Ratio (PSNR) was calculated as $\infty$ dB, confirming lossless recovery.
• Security: Experiments demonstrated that reconstructing the image with fewer than $n$ shares (e.g., 2 or 3 shares in a (4,4) scheme) resulted in random, meaningless images, confirming the security condition.
• Decodability: The generated CQR shares were successfully decoded by a standard WeChat decoder (version 8.0.54). The study found that a scaling factor (min) of at least 2 is required for the shares to be decodable, with a factor of 5 or higher recommended to ensure the shares appear visually natural to the Human Visual System (HVS).

Significance and Claims
The paper claims that the CQRVSS scheme offers a significant advancement over existing methods by solving the trade-off between share meaningfulness and recovery quality.

• Enhanced Security: By replacing meaningless noise with recognizable QR codes, the scheme mitigates the risk of targeted attacks on suspicious data packets.
• Practicality: The use of standard QR code infrastructure and simple XOR operations makes the scheme feasible for real-world deployment in public networks.
• Applicability: The ability to perform lossless recovery makes the scheme suitable for sensitive, detail-dependent images (e.g., ID cards, health records, confidential organizational data) that cannot tolerate the artifacts common in halftone-based or fuzzy-grid VSS schemes.

The authors conclude that while the current implementation is an (n, n)-threshold scheme, the framework provides a robust foundation for future work extending to more flexible (k, n)-threshold configurations.