PDE based scheme for multi-modal medical image watermarking

Area of Science:

• Medical Imaging
• Computer Science
• Digital Security

Background:

• Digital watermarking is crucial for protecting intellectual property rights of medical images exchanged online.
• Ensuring the integrity and authenticity of medical images is a significant challenge.
• Existing Region of Interest (ROI)/Region of Non-Interest (RONI) watermarking schemes have limitations in efficiency and applicability.

Purpose of the Study:

• To develop an automatic Region of Interest (ROI)/Region of Non-Interest (RONI) selection algorithm for digital watermarking.
• To address the limitations of manual or semi-automatic ROI/RONI selection in medical imaging.
• To improve the robustness, transparency, and authenticity of watermarked medical images.

Main Methods:

• A novel automatic ROI/RONI selection algorithm based on the Partial Differential Equation (PDE) method.
• Embedding watermarks in selected regions to ensure copyright protection.
• Evaluation using various medical imaging modalities including MRI, Ultrasound, CT, and X-ray.

Main Results:

• The proposed method achieves high imperceptibility, with Peak Signal-to-Noise Ratio (PSNR) values ranging from 60.13 to 94.75 dB.
• Demonstrates good robustness against various attacks, with Normalized Correlation (NC) values >= 0.998.
• Effectively detects and recovers tampered regions in medical images across different modalities.

Conclusions:

• The developed watermarking technique enhances robustness against attacks and improves transparency.
• Provides superior authenticity and integrity protection for digital medical image content.
• Minimizes complexity associated with ROI/RONI selection, making it efficient for large-scale applications.