Low latency and generalizable dynamic MRI via L+S alternating gradient descent and minimization
- 1Department of Electrical and Computer Engineering Iowa State University Ames, Iowa USA
- 2Roy J. Carver Department of Biomedical Engineering University of Iowa Iowa City, Iowa USA
Abstract
Background: Dynamic MRI reconstruction often requires accurate, fast, and low‐latency algorithms that generalize across multiple applications, sampling schemes, and rates, without problem‐specific parameter tuning. Existing methods either lack generalizability or require extensive parameter adjustment.
Purpose: To develop generalizable, low‐latency MRI reconstruction algorithms based on low‐rank (LR) and low‐rank + sparse (L+S) models that are accurate, fast, and require minimal parameter tuning across diverse dynamic MRI settings.
Methods: We propose a novel Alternating Gradient Descent and Minimization (AltGDmin) approach for L+S matrix recovery from column‐wise undersampled measurements, and extend it to a 3‐level hierarchical L+S model for dynamic MRI. A few‐shot (FS) learning modification (AGM‐L+S‐FS) is introduced to enable near low‐latency reconstruction with per‐frame processing time comparable to or faster than scan time. A simpler LR‐only variant is also evaluated.
Results: The proposed algorithms were validated on 6 prospective and 12 retrospective dynamic MRI datasets, covering cardiac, speech, and abdominal applications, with Cartesian, pseudo‐radial, radial, and spiral sampling schemes. Experiments demonstrate accurate reconstruction, generalizability across datasets and sampling rates, and low‐latency performance. Detailed comparisons with state‐of‐the‐art LR, L+S, low‐latency, and unsupervised deep learning methods show superior or comparable performance.
Conclusions: Our AGM‐L+S‐FS approach provide generalizable, fast, and low‐latency MRI reconstruction across multiple dynamic MRI applications and sampling schemes, without the need for problem‐specific parameter tuning.
Related
- Chaos and dynamic instabilities of FGM micro-pipes with internal fluid flow via strain gradient theory
- Evolution of colorectal cancer screening technologies and health economic evaluation: optimization strategies based on an economic gradient and dynamic risk stratification
- An Explainable Disease-Aware Intelligent Histopathological Representation Learning Framework for Generalizable Breast Cancer Classification
- A generalizable pipeline for specialty-curated postmarket device surveillance: a reference implementation in neurosurgery
- HrasG12V induces follicular thyroid cancer with attenuated MAPK activation and increased latency compared to KrasG12D
BibTeX
@article{babu2026latency,
title = {{Low latency and generalizable dynamic MRI via L+S alternating gradient descent and minimization}},
author = {Silpa Babu and Sajan Goud Lingala and Namrata Vaswani},
journal = {Medical Physics},
year = {2026},
month = oct,
doi = {10.1002/mp.70713},
url = {https://doi.org/10.1002/mp.70713},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Low latency and generalizable dynamic MRI via L+S alternating gradient descent and minimization
AU - Silpa Babu
AU - Sajan Goud Lingala
AU - Namrata Vaswani
PY - 2026
DA - 2026/10/27
JO - Medical Physics
DO - 10.1002/mp.70713
UR - https://doi.org/10.1002/mp.70713
AB - Background: Dynamic MRI reconstruction often requires accurate, fast, and low‐latency algorithms that generalize across multiple applications, sampling schemes, and rates, without problem‐specific parameter tuning. Existing methods either lack generalizability or require extensive parameter adjustment. Purpose: To develop generalizable, low‐latency MRI reconstruction algorithms based on low‐rank (LR) and low‐rank + sparse (L+S) models that are accurate, fast, and require minimal parameter tuning across diverse dynamic MRI settings. Methods: We propose a novel Alternating Gradient Descent and Minimization (AltGDmin) approach for L+S matrix recovery from column‐wise undersampled measurements, and extend it to a 3‐level hierarchical L+S model for dynamic MRI. A few‐shot (FS) learning modification (AGM‐L+S‐FS) is introduced to enable near low‐latency reconstruction with per‐frame processing time comparable to or faster than scan time. A simpler LR‐only variant is also evaluated. Results: The proposed algorithms were validated on 6 prospective and 12 retrospective dynamic MRI datasets, covering cardiac, speech, and abdominal applications, with Cartesian, pseudo‐radial, radial, and spiral sampling schemes. Experiments demonstrate accurate reconstruction, generalizability across datasets and sampling rates, and low‐latency performance. Detailed comparisons with state‐of‐the‐art LR, L+S, low‐latency, and unsupervised deep learning methods show superior or comparable performance. Conclusions: Our AGM‐L+S‐FS approach provide generalizable, fast, and low‐latency MRI reconstruction across multiple dynamic MRI applications and sampling schemes, without the need for problem‐specific parameter tuning.
SN - 0094-2405
SN - 2473-4209
ER -