A cardiac MRI motion artifact reduction method based on edge enhancement network.

Area of Science:

• Medical Imaging
• Image Processing
• Biomedical Engineering

Background:

• Cardiac MRI (magnetic resonance imaging) acquisition is lengthy, often leading to patient motion artifacts.
• Image artifacts in cardiac MRI can obscure pathological details, hindering accurate diagnosis.
• Clear texture edges in MR images are vital for identifying cardiac abnormalities.

Purpose of the Study:

• To develop a novel method for reducing motion artifacts in cardiac MRI.
• To enhance the clarity of texture edges in cardiac MRI for improved pathological diagnosis.
• To propose an edge enhancement network for de-artifacting cardiac MRI.

Main Methods:

• Utilized a four-plane normal vector adaptive fractional differential mask for edge feature extraction, reducing noise.
• Adaptive fractional order selection based on image gradients and Gaussian curvature entropy.
• Developed an edge enhancement network incorporating edge fusion feature extraction and transformer networks.
• Designed attention mechanisms within the transformer network to focus on blurred image edges.

Main Results:

• The proposed method significantly reduced motion artifacts in cardiac MRI.
• Achieved higher peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) compared to existing methods.
• Resulting de-artifacted images exhibited clear and well-defined texture edges.

Conclusions:

• The edge enhancement network effectively reduces motion artifacts in cardiac MRI.
• The method preserves and enhances critical texture edge information for diagnostic purposes.
• This approach offers a promising solution for improving the quality and diagnostic utility of cardiac MRI.