ACCELERATING CARDIOVASCULAR IMAGING BY EXPLOITING REGIONAL LOW-RANK STRUCTURE VIA GROUP SPARSITY

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Magnetic Resonance Imaging

Background:

• Sparse sampling in (k, t)-space is crucial for cardiac MRI.
• Partial separability (PS) and spatial-spectral sparsity enable high-quality reconstruction from undersampled data.

Purpose of the Study:

• To develop a more flexible method for controlling PS-induced low-rank constraints in cardiac MRI.
• To introduce a novel algorithm for solving the (1,2)-norm regularized inverse problem in undersampled (k, t)-space MRI.

Main Methods:

• Utilizing group-sparse regularization for enhanced control over low-rank constraints.
• Developing and applying a novel algorithm to address the (1,2)-norm regularized inverse problem.
• Reconstructing images from highly undersampled (k, t)-space data.

Main Results:

• Demonstrated improved performance in reconstructing cardiac MRI data.
• Showcased the effectiveness of group-sparse regularization for flexible constraint control.
• Validated the novel algorithm's ability to handle the inverse problem.

Conclusions:

• The proposed method offers enhanced control and improved image reconstruction quality for cardiac MRI.
• The novel algorithm effectively solves the (1,2)-norm regularized inverse problem in sparse MRI.
• This approach advances high-quality image reconstruction from undersampled (k, t)-space data.