This project focuses on the development of open-source, rapid, and memory-efficient software tools for radiofrequency (RF) coil modeling in magnetic resonance imaging (MRI), with a focus on ultra-high field (UHF) MRI applications at field strength higher than 3 tesla.

While UHF MRI offers superior anatomical detail for early disease diagnosis, it also introduces challenges due to the shorter RF wavelengths relative to body dimensions. These conditions can lead to complex interference patterns that cause local signal dropouts and RF energy deposition hotspots, which deteriorate image quality and raise patient safety concerns.

To address these issues, we are developing novel electromagnetic (EM) simulation tools based on domain decomposition techniques and integral equation methods (wire, surface, and volume formulations). A key innovation is the integration of tensor decomposition techniques to construct compressed, low-memory representations of EM interactions, thus enabling significant acceleration of simulations. Additionally, we are developing an RF circuit co-simulator that automatizes coil tuning, matching, decoupling, and detuning via a pipeline of particle swarm optimizations.

Our simulation tools are the first to leverage tensor decompositions in RF coil design for MRI. By removing licensing costs and substantially reducing computation times, our software aims to offer accessible, high-performance tools for researchers and engineers developing next-generation MRI hardware.