Fast MR image reconstruction for partially parallel imaging with arbitrary k-space trajectories

Area of Science:

• Medical Imaging
• Magnetic Resonance Imaging
• Computational Imaging

Background:

• Accelerating Magnetic Resonance (MR) imaging acquisition and reconstruction is vital for clinical practice.
• Parallel MR imaging techniques enhance speed but require efficient reconstruction algorithms.
• SENSE (Sensitivity Encoding) is a key parallel imaging method.

Purpose of the Study:

• To develop a fast and efficient reconstruction algorithm for SENSE in partially parallel MR imaging.
• To handle arbitrary k-space trajectories for greater flexibility in MR imaging.
• To improve both computational efficiency and reconstruction quality.

Main Methods:

• The proposed method combines variable splitting, penalty techniques, and optimal gradient methods.
• It reformulates the SENSE model with sparsity regularization into an unconstrained minimization problem.
• Alternating minimizations solve total variation/wavelet denoising and linear inversion problems.

Main Results:

• The algorithm significantly improves computational efficiency compared to existing parallel imaging methods.
• It achieves state-of-the-art reconstruction quality.
• The method effectively handles arbitrary k-space trajectories.

Conclusions:

• The developed fast reconstruction algorithm offers a significant advancement for SENSE in parallel MR imaging.
• This method enhances clinical applicability by improving speed and maintaining high image quality.
• It provides a robust solution for MR imaging with complex k-space trajectories.