Motion Compensated Dynamic MRI Reconstruction With Local Affine Optical Flow Estimation.

Area of Science:

• Medical Imaging
• Image Reconstruction
• Magnetic Resonance Imaging

Background:

• Dynamic MRI (DMRI) is crucial for visualizing physiological processes.
• Motion artifacts significantly degrade DMRI quality.
• Existing reconstruction methods struggle with accurate motion compensation.

Purpose of the Study:

• To develop a novel framework for DMRI reconstruction with integrated motion compensation (MC).
• To jointly reconstruct DMRI sequences and estimate interframe motion vectors.
• To refine DMRI reconstruction using estimated motion fields across multiple scales.

Main Methods:

• Combined intensity-based optical flow constraints with compressed sensing for joint reconstruction and motion estimation.
• Employed a coarse-to-fine multi-scale resolution strategy for motion field updates.
• Utilized a primal-dual algorithm with linesearch to solve the optimization problem, accommodating various priors like sparsity, low rank, and total variation.

Main Results:

• Successfully reconstructed DMRI sequences with effective motion compensation.
• Demonstrated significant improvement in DMRI reconstruction quality compared to state-of-the-art methods.
• Validated the framework's ability to handle diverse prior information for enhanced reconstruction.

Conclusions:

• The proposed framework offers a robust solution for high-quality DMRI reconstruction.
• Joint reconstruction and motion compensation significantly reduce motion artifacts.
• The method's flexibility in incorporating prior information enhances its applicability across various DMRI scenarios.