Reference guided CS-MRI with gradient orientation priors.

Area of Science:

• Medical Imaging
• Signal Processing
• Computational Science

Background:

• Compressed Sensing (CS) is a framework for Magnetic Resonance Imaging (MRI) reconstruction from under-sampled k-space data.
• High acceleration in CS-MRI can lead to severe aliasing artifacts, compromising image quality.
• Reference images, easily acquired in many MRI applications, offer potential for artifact reduction.

Purpose of the Study:

• To propose a novel reference-guided Compressed Sensing MRI reconstruction method.
• To leverage gradient orientation priors (GOP) for enhanced reconstruction accuracy.
• To address aliasing artifacts in high-acceleration MRI scans.

Main Methods:

• A novel reference-guided CS-MRI reconstruction method utilizing gradient orientation priors (GOP) was developed.
• The method regularizes the tangent vector in the target image to be perpendicular to the normal vector in the reference image.
• Validation was performed using multi-scan experimental data.

Main Results:

• The proposed GOP-based method effectively reduces aliasing artifacts in high-acceleration CS-MRI.
• The technique demonstrates high-speed imaging capabilities.
• High-quality MR images were achieved, outperforming existing methods.

Conclusions:

• The novel reference-guided CS-MRI reconstruction method using GOP offers a significant advancement.
• This approach provides a viable solution for high-quality, accelerated MRI acquisition.
• The method shows promise for improving clinical MRI workflows through faster scan times and reduced artifacts.