MRI Simulation-based evaluation of an efficient under-sampling approach.

Area of Science:

• Medical Imaging
• Signal Processing
• Applied Physics

Background:

• Compressive sampling (CS) is vital for reconstructing sparse signals in magnetic resonance imaging (MRI).
• Traditional random undersampling in MRI's phase-encoding (k_y) dimension is limited.
• Exploiting k-space origin data is crucial for efficient MRI reconstruction.

Purpose of the Study:

• To introduce a novel hybrid undersampling strategy for 2D Cartesian k-space MRI.
• To improve MRI reconstruction accuracy and efficiency using compressive sampling.
• To reduce computational complexity in MRI reconstruction.

Main Methods:

• Proposed a hybrid undersampling approach dividing k_y measurements into 70% random and 30% definite sampling near k-space origin.
• Utilized a simple and efficient CS-based simulation for MRI reconstruction.
• Compared the hybrid scheme against traditional random undersampling.

Main Results:

• The hybrid scheme significantly improved average error (up to 55%) and universal image quality index (up to 77%) at lower undersampling ratios (r).
• Demonstrated substantial gains compared to the traditional random undersampling method.
• Achieved these improvements using a computationally efficient CS method.

Conclusions:

• The proposed hybrid undersampling approach offers a simple, efficient, and effective method for MRI reconstruction improvement.
• This technique enhances image quality and reduces errors in MRI scans.
• Findings are valuable for developing faster MRI acquisition techniques and reducing scan times.