DATA-DRIVEN IMAGING GEOMETRIC RECOVERY OF ULTRAHIGH RESOLUTION ROBOTIC MICRO-CT FOR IN-VIVO AND OTHER APPLICATIONS

Area of Science:

• Medical Imaging
• Robotics
• Artificial Intelligence

Background:

• Carotid plaques require high-resolution imaging for accurate diagnosis.
• Existing computed tomography (CT) methods face challenges with geometric artifacts in localized scans.

Purpose of the Study:

• To introduce an ultrahigh-resolution robotic micro-CT system for carotid plaque imaging.
• To develop a data-driven method for correcting geometric errors in interior CT scans.

Main Methods:

• Utilized a robotic micro-CT design with advanced detectors and machine learning.
• Implemented a data-driven geometry estimation using normalized cross-correlation to address projection truncation.
• Validated the method on robotic CT scans of a mouse and a micro-CT phantom.

Main Results:

• Achieved ultrahigh resolution of 50 micrometers for localized imaging.
• The proposed geometry correction method significantly reduced artifacts.
• Resulting images exhibited sharper details compared to previous correction techniques.

Conclusions:

• The robotic micro-CT system offers superior resolution for carotid plaque imaging.
• The data-driven geometry correction method effectively enhances image quality by minimizing artifacts.
• This technology shows promise for improved diagnosis and research of vascular diseases.