Distribution matching with subset-k-space embedding for multi-contrast MRI reconstruction.

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging (MRI)
• Image Reconstruction

Background:

• Multi-contrast MRI (MC-MRI) is crucial for diagnostics but suffers from motion artifacts due to long acquisition times.
• Current methods reconstruct MC-MRI from partial k-space data, leveraging inter-contrast redundancy.
• Exploiting cross-contrast information offers a more effective path for accurate MC-MRI reconstruction.

Purpose of the Study:

• To develop a novel MC-MRI reconstruction method integrating subset-k-space distribution and high-dimensional global priors.
• To enhance reconstruction accuracy and efficiency in MC-MRI.

Main Methods:

• A two-stage approach: 1) Individual k-space decomposition and subset-k-space construction using distribution matching.
• 2) Global prior embedding to constrain the diffusion model within a high-dimensional space, referencing the reconstructed contrast.

Main Results:

• The proposed DMSE method demonstrates excellent capability in preserving details and achieving accurate MC-MRI reconstruction.
• Empirical evaluations across diverse datasets validate the method's performance.

Conclusions:

• DMSE integrates subset-k-space and high-dimensional global priors for guided MC-MRI reconstruction.
• The model effectively reduces noise and aliasing artifacts by leveraging guidance contrasts and self-constrained information.
• Comparative studies show DMSE outperforms existing methods in quantitative and qualitative evaluations.