Projection onto Epigraph Sets for Rapid Self-Tuning Compressed Sensing MRI

Area of Science:

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

Background:

• Compressed Sensing (CS) in MRI relies on image sparsity for reconstructing undersampled data.
• Regularization parameters are crucial for balancing sparsity and data fidelity but are dataset-specific, limiting generalization.
• Current heuristic parameter selection methods often perform poorly on new datasets.

Purpose of the Study:

• To develop a computationally efficient, self-tuning Compressed Sensing (CS) method for Magnetic Resonance Imaging (MRI) reconstruction.
• To address the limitations of traditional parameter selection methods that are time-consuming and dataset-dependent.
• To improve the speed of CS MRI reconstruction while maintaining high image quality.

Main Methods:

• Proposed a novel self-tuning CS method utilizing efficient projections onto epigraph sets of L1 and Total Variation norms.
• Integrated simultaneous parameter selection and regularization within the reconstruction process.
• Validated the method on in vivo data across balanced steady-state free precession, time-of-flight, and T1-weighted imaging sequences.

Main Results:

• The proposed self-tuning CS method achieved an order of magnitude improvement in computational efficiency compared to existing line-search techniques.
• Near-optimal parameter selection was maintained, ensuring high-quality image reconstruction.
• Demonstrated successful application across various MRI pulse sequences.

Conclusions:

• The novel self-tuning CS method offers a significant advancement in MRI reconstruction efficiency.
• This approach overcomes the limitations of traditional parameter tuning, providing a robust and faster alternative.
• The method holds promise for accelerating clinical MRI workflows and improving patient experience.