A New Evaluation Method for Medical Image Information Hiding Techniques

Area of Science:

• Medical Imaging
• Radiology
• Information Security

Background:

• Medical image scans and electronic medical records (EMR) require robust security for teleradiology and remote healthcare.
• Information hiding (IH) techniques are increasingly relevant for securing data transmitted over open networks.
• Current evaluations of IH algorithms for medical images primarily use signal processing parameters, potentially overlooking diagnostic features.

Purpose of the Study:

• To re-interpret existing evaluation parameters for IH algorithms in medical imaging.
• To propose a new framework for the dynamic selection of medical image IH security algorithms (watermarking and steganography).
• To incorporate criteria reflecting medical statistics crucial for patient diagnosis and monitoring.

Main Methods:

• Developed a novel evaluation framework incorporating medical statistical criteria.
• Validated the framework on the Pneumonia Chest Xray dataset.
• Tested three selected IH algorithms focusing on privacy and tamper detection.

Main Results:

• The new framework allows for a more comprehensive evaluation of IH algorithms beyond traditional signal processing metrics.
• Demonstrated the framework's ability to select appropriate IH algorithms that preserve diagnostic integrity.
• Successfully applied the framework to assess security measures for chest X-ray images.

Conclusions:

• The proposed framework enhances the security of medical images in teleradiology by ensuring diagnostic features are maintained.
• This approach facilitates the dynamic selection of IH algorithms tailored to specific medical diagnostic needs.
• The study provides a validated method for assessing the suitability of watermarking and steganography for secure medical data transmission.