Halftone-image security improving using overall minimal-error searching.

Area of Science:

• Digital Watermarking
• Image Processing
• Information Security

Background:

• Achieving high embedding capacity, data integrity, and visual quality in image-based data hiding remains challenging.
• Existing methods often compromise one aspect when optimizing for others.

Purpose of the Study:

• To develop a novel data hiding method, overall minimal-error searching (OMES), that balances high embedding capacity, perfect data extraction, and superior image quality.
• To enhance security through secret sharing across multiple halftone images.
• To improve robustness against print-and-scan attacks.

Main Methods:

• Implemented overall minimal-error searching (OMES) using a trained substitution table (S-Table) to modify halftone values across multiple host images.
• Employed secret sharing to distribute watermarks, requiring all marked images for extraction.
• Developed error-diffused-based and least-mean-square-based encoders for speed and quality optimization.
• Introduced a compensation correction procedure to counter print-and-scan attack distortions.

Main Results:

• The proposed OMES method successfully achieves high embedding capacity and 100% data extraction while maintaining good image quality.
• Experimental results demonstrate the effectiveness of the S-Table optimization in preserving visual fidelity.
• The method shows resilience against common print-and-scan attack artifacts like rotation and zooming.

Conclusions:

• The OMES method offers a robust solution for secure image-based data hiding, addressing the trade-offs between capacity, extraction, and quality.
• The integration of secret sharing and optimized S-Tables provides enhanced security and visual imperceptibility.
• The developed approach is suitable for applications involving printing and scanning, ensuring data integrity and quality.