arXiv subjects

Charlotte R. Sappo

Publications and source records attributed to Charlotte R. Sappo.

4 recordsLinked to original sources

arXiv2026

PGDM-MRSRGAN: Physics-Guided Degradation Model with an SRGAN Framework for Magnetic Resonance Image Super-Resolution: Applications in Low-Field MRI

Magnetic Resonance Imaging (MRI) often suffers from low signal-to-noise ratio (SNR) and limited spatial resolution, which compromise clinical precision. This study aims to address these challenges by developing a physics-guided degradation model (PGDM) and a novel deep learning framework, Magnetic Resonance Super-Resolution GAN Imaging (MRSRGAN), for improved super-resolution reconstruction of MRI images. The proposed approach consists of two stages: (1) a physics-guided degradation model that simulates low-field MRI conditions by incorporating artifacts such as blur, B0 inhomogeneity, chemical shift effects, down-sampling, and noise to generate paired low-resolution (LR) and high-resolution (HR) training datasets; and (2) an MRSRGAN-based reconstruction network utilizing a hybrid loss function, including modified spectral angle mapper (MSAM) and sharpened ground truth images based on de-convolution of slice profile-based kernels to restore high-SNR and high-resolution images. Experimental validation demonstrates that the proposed method achieves superior spatial resolution, enhanced SNR, improved peak signal-to-noise ratio (PSNR), higher structural similarity index (SSIM), reduced MSAM error, and better non-reference image quality evaluation (NIQE) metrics. Furthermore, it effectively reduces artifacts and demonstrates robustness on real MRI datasets. The MRSRGAN framework, guided by the physics-informed degradation model, provides a significant improvement in MRI image quality, enhancing spatial resolution and diagnostic accuracy. Its demonstrated robustness and effectiveness on real MRI datasets highlight its potential to resolution enhancement for precise diagnostic imaging.

eess.IV↗

arXiv2026

LoFi RADIO: A Distilled In-Domain Backbone Applied for Artifact-Severity Grading of Ultra-Low-Field Neonatal Brain MR

Ultra-low-field MRI makes neonatal brain imaging deploy- able in low-resource settings, but its low SNR, lack of shielding, and long scan duration make it especially prone to acquisition artifacts, motivating automated quality control. We address the LISA 2026 Task 1a challenge: multi-label severity grading (0/1/2) of seven common image artifacts on ULF T2 weighted volumes. We identify that a number of backbones may be successfully paired with a classification MLP, but that no single backbone is uniformly best across artifacts. To improve performance, we evaluate routing complementary foundation model teachers through a per-artifact gate, as well as distilling the teachers into a single in-domain ViT-S student (LoFi RADIO) over an unlabeled low-field MRI corpus. Both of these strategies improve the weighted composite. The distilled backbone matches or exceeds the gate and has the added advantage of not requiring deployment of multiple large foundation models at infer- ence.

eess.IV↗

arXiv2024

A Minibatch Alternating Projections Algorithm for Robust and Efficient Magnitude Least-Squares RF Pulse Design in MRI

A magnitude-least-squares radiofrequency pulse design algorithm is reported which uses interleaved exact and stochastically-generated inexact updates to escape local minima and find low-cost solutions. Inexact updates are performed using a small randomly selected minibatch of the available B1+ measurements to update RF pulse weights, which perturbs the sequence of alternating projections. Applications to RF shimming, parallel transmit spokes RF pulse design, and spectral-spatial RF pulse design are considered. Numerical and simulation studies characterized the optimal minibatch size, which was found to consistently produce lower power and lower RMSE solutions across subjects, coil geometries, B1+ resolutions and orientations. The method was validated in-vivo at 7 Tesla and produced improvements in image quality in a slice-by-slice RF-shimmed imaging sequence. Compared to conventional methods, the pulse design method can more robustly design RF pulses that correct for B1+ inhomogeneities at ultra-high field strengths, and enable pulse designs to be completed with increased computational efficiency

physics.med-ph↗

arXiv2020

A Portable Brain MRI Scanner for Underserved Settings and Point-Of-Care Imaging

Access to and availability of MRI scanners is typically limited by their cost, siting and infrastructure requirements. This precludes MRI diagnostics, the reference standard for neurological assessment, in patients who cannot be transported to specialized scanner suites. This includes patients who are critically ill and unstable, and patients located in low-resource settings. The scanner design presented here aims to extend the reach of MRI by substantially reducing these limitations. Our goal is to shift the cost-benefit calculation for MRI toward more frequent and varied use, including improved accessibility worldwide and point of care operation. Here, we describe a portable brain MRI scanner using a compact, lightweight permanent magnet, with a built-in readout field gradient. Our low-field (80 mT) Halbach cylinder design of rare-earth permanent magnets results in a 122 kg magnet with minimal stray-field, requiring neither cryogenics nor external power. The built-in magnetic field gradient reduces reliance on high-power gradient drivers, which not only lowers overall system power and cooling requirements, but also reduces acoustic noise. Imperfections in the encoding fields are mitigated with a generalized iterative image reconstruction technique, that uses prior characterization of the field patterns. Our system was validated using T1, T2 and proton density weighted in vivo brain images with a spatial resolution of 2.2 x 1.3 x 6.8 mm$^3$.

eess.IV↗