TranSMS: Transformers for Super-Resolution Calibration in Magnetic Particle Imaging

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Nanotechnology

Background:

• Magnetic particle imaging (MPI) provides high-resolution imaging of magnetic nanoparticles (MNPs).
• MPI calibration using system matrix (SM) measurements is crucial for accurate MNP distribution reconstruction but is time-consuming.
• Existing calibration methods struggle with system variations, necessitating frequent recalibration.

Purpose of the Study:

• To develop a novel deep learning approach for accelerated MPI calibration.
• To improve the efficiency and accuracy of system matrix (SM) recovery in MPI.
• To enable faster and more reliable MNP imaging through reduced calibration times.

Main Methods:

• Introduced TranSMS, a deep learning model based on Transformers for super-resolution of low-resolution SM measurements.
• Utilized large MNP samples for efficient low-resolution SM acquisition with improved signal-to-noise ratio.
• Integrated a vision transformer module, a dense convolutional module, and a data-consistency module within the TranSMS framework.

Main Results:

• TranSMS achieved significant improvements in SM recovery and MPI reconstruction accuracy.
• Demonstrated up to 64-fold acceleration in two-dimensional MPI calibration and imaging.
• Validated the approach using both simulated and experimental MPI data.

Conclusions:

• TranSMS offers a highly effective solution for accelerating MPI calibration.
• The deep learning approach significantly reduces calibration time while maintaining or improving image reconstruction quality.
• This advancement has the potential to broaden the applicability of MPI in various fields.