Medical Image Tamper Detection Based on Passive Image Authentication

Area of Science:

• Medical Imaging
• Computer Science
• Digital Forensics

Background:

• Telemedicine relies on transmitting medical images, raising concerns about data integrity.
• Existing watermarking methods for medical image authentication can degrade image quality and lead to misdiagnosis.
• Passive image authentication is crucial for verifying medical image authenticity without altering the original image.

Purpose of the Study:

• To develop a passive image authentication mechanism for detecting tampered regions in medical images.
• To improve the robustness of keypoint-based authentication methods against various image manipulations and attacks.
• To ensure the integrity of medical images used in telemedicine and telediagnosis.

Main Methods:

• Utilized Local Binary Pattern Rotation Invariant (LBPROT) to extract structural texture information.
• Employed Scale Invariant Feature Transform (SIFT) for keypoint extraction on texture images.
• Developed a keypoint matching approach to identify and locate tampered regions within medical images.

Main Results:

• The proposed method successfully detected tampered regions in medical images.
• Authentication accuracy was maintained even when forged regions were scaled or rotated.
• The method proved effective against common image attacks like Gaussian blurring and additive white Gaussian noise.

Conclusions:

• The LBPROT-enhanced keypoint-based passive authentication method improves medical image integrity verification.
• This technique offers a robust solution for detecting tampering in medical images, crucial for secure telemedicine.
• The findings contribute to enhancing the reliability of digital medical records and diagnostic processes.