A survey of GPU-based acceleration techniques in MRI reconstructions.

Area of Science:

• Medical Imaging
• Computational Science
• Computer Engineering

Background:

• Magnetic resonance imaging (MRI) reconstruction is computationally intensive.
• Clinical applications demand rapid image processing for diagnosis and treatment.
• Graphics processing units (GPUs) offer high-performance parallel computing capabilities.

Purpose of the Study:

• To survey image reconstruction schemes utilizing GPU computing in MRI.
• To provide a reference for researchers in the MRI community.
• To highlight the growing importance of GPUs, especially with deep learning integration.

Main Methods:

• Review of existing literature on GPU-accelerated MRI image reconstruction.
• Analysis of computational performance and efficiency of GPU-based methods.
• Categorization of reconstruction techniques based on GPU implementation.

Main Results:

• GPUs enable faster and more efficient MRI image reconstruction.
• GPU computing is becoming essential for complex MRI reconstruction tasks.
• The integration of deep learning further enhances GPU utility in MRI.

Conclusions:

• GPU computing is a critical technology for modern MRI image reconstruction.
• This survey provides a valuable resource for researchers exploring GPU acceleration in MRI.
• Future MRI advancements will likely leverage GPU capabilities, particularly with AI.