A reversible-zero watermarking scheme for medical images

Area of Science:

• Medical Imaging
• Digital Watermarking
• Cybersecurity

Background:

• Telemedicine relies heavily on medical images, necessitating robust methods for authenticity and copyright protection.
• Existing authentication techniques often lack sufficient resilience against unintentional attacks.
• A gap exists in research regarding the robustness of medical image authentication against accidental data corruption.

Purpose of the Study:

• To propose a novel Reversible-Zero Watermarking approach for medical images.
• To enhance the resilience of medical image authentication against unintentional attacks.
• To ensure copyright protection and authenticity in telemedicine applications.

Main Methods:

• A hybrid approach combining zero-watermarking and reversible watermarking techniques.
• Utilizing VGG19-based feature extraction for ownership establishment.
• Employing Discrete Wavelet Transform, Integer Wavelet Transform, and Difference Expansion for reversible embedding.

Main Results:

• The proposed method demonstrates significant improvements over existing techniques.
• High Normalization Correlation (NC) values (around 0.9) and low Bit Error Rate (BER) values (close to 0) were achieved under various attacks.
• The watermarking system proved to be imperceptible, distinguishable, and robust.

Conclusions:

• The Reversible-Zero Watermarking approach offers a robust and secure solution for medical image authentication in telemedicine.
• The system provides persistent verification independent of third-party storage.
• This method is well-suited for safeguarding medical image integrity and ownership.