Non-iterative image reconstruction from sparse magnetic resonance imaging radial data without priors

Area of Science:

• Medical Imaging
• Biophysics
• Signal Processing

Background:

• Current state-of-the-art magnetic resonance imaging (MRI) reconstruction relies on iterative compressed sensing algorithms.
• These iterative methods optimize objective functions incorporating spatial and/or temporal constraints.
• Compressed sensing (CS) algorithms can be computationally intensive and time-consuming.

Purpose of the Study:

• To propose a novel non-iterative algorithm for image reconstruction from under-sampled k-space data.
• To compare the proposed non-iterative method against state-of-the-art iterative compressed sensing techniques.
• To demonstrate the feasibility and efficacy of the new reconstruction approach.

Main Methods:

• A non-iterative algorithm was developed to estimate un-measured k-space data.
• The estimated data was used with the filtered backprojection (FBP) algorithm for image reconstruction.
• The proposed method was validated using a patient MRI study.
• Performance was compared against iterative total-variation (TV) compressed sensing reconstruction.

Main Results:

• The proposed non-iterative method successfully reconstructed images from under-sampled k-space data.
• Feasibility was demonstrated in a clinical patient MRI study.
• The non-iterative approach showed comparable or superior performance to iterative TV-based compressed sensing in certain aspects.
• This suggests a potential for faster image reconstruction without significant loss of image quality.

Conclusions:

• A non-iterative algorithm offers a viable alternative to iterative compressed sensing for MRI reconstruction.
• The proposed method provides efficient image reconstruction from under-sampled k-space data.
• This approach has the potential to accelerate MRI acquisition and reconstruction workflows.