Low-rank plus sparse joint smoothing model based on tensor singular value decomposition for dynamic MRI...

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Image Reconstruction

Background:

• Dynamic magnetic resonance imaging (DMRI) is crucial for medical diagnosis but is limited by long acquisition times.
• Accelerating DMRI acquisition is essential for broader clinical applications and improved patient comfort.

Purpose of the Study:

• To develop an advanced reconstruction method for highly under-sampled DMRI data.
• To enhance the quality and detail of dynamic MR images while reducing scan duration.

Main Methods:

• A low-rank plus sparse tensor (ℒ+S) model was proposed, leveraging tensor singular value decomposition (T-SVD).
• The model exploits spatiotemporal correlations inherent in DMRI data.
• Joint tensor total variation (TTV) constraints were incorporated to preserve global structure and enhance reconstruction details.

Main Results:

• The proposed method successfully reconstructed dynamic MR images from highly under-sampled k-t space data.
• Experimental results on dynamic cardiac datasets demonstrated superior performance compared to existing state-of-the-art methods.
• The approach effectively improved reconstruction quality and preserved image details.

Conclusions:

• The proposed low-rank plus sparse joint smoothing model effectively accelerates DMRI acquisition.
• This method offers a promising solution for faster and more detailed dynamic MR imaging.
• The technique shows significant potential for clinical translation in various DMRI applications.