An IPVO-based reversible data hiding scheme using floating predictors

Area of Science:

• Computer Science
• Information Security
• Digital Image Processing

Background:

• Reversible data hiding (RDH) allows secret data embedding without permanent image distortion.
• Existing RDH schemes, like Peng et al.'s, face challenges in minimizing visual changes and pixel shifts.
• Improved Pixel-Value-Ordering (IPVO) and prediction-error expansion are key techniques in RDH.

Purpose of the Study:

• To optimize a high-fidelity reversible data hiding scheme.
• To reduce the pixel shift rate and visual impact on the carrier image.
• To improve the performance of RDH schemes, especially for rough images.

Main Methods:

• Optimizing Peng et al.'s IPVO-based RDH scheme.
• Selecting appropriate prediction-errors for data embedding to minimize pixel shifts.
• Extending the IPVO-based RDH scheme with a floating predictor.

Main Results:

• The proposed scheme achieves superior performance compared to state-of-the-art methods at the same embedding capacity.
• Demonstrated significant reduction in pixel shifts, especially for rough images.
• Successfully reduced visual changes in the carrier image.

Conclusions:

• The optimized IPVO-based RDH scheme offers improved high-fidelity data hiding.
• The selection of prediction-errors is crucial for minimizing image modification.
• The proposed method provides a robust solution for secure data embedding in digital images.