Denoising very low-field magnetic resonance images using native noise modeling
Bibliographic record
Abstract
Low-field MRI is gaining interest, especially in low-resource settings, due to its low cost, portability, small footprint, and low power consumption. However, it suffers from significant noise, limiting its clinical utility. This study introduces native noise denoising (NND), which leverages the inherent noise characteristics of the acquired low-field data. By obtaining the noise characteristics from corner patches of low-field images, we iteratively added similar noise to high-field images to create a paired noisy-clean dataset. A U-Net based denoising autoencoder was trained on this dataset and evaluated on three low-field datasets: the M4Raw dataset (0.3T), in vivo brain MRI (0.05T), and phantom images (0.05T). The NND approach demonstrated improvements in signal-to-noise ratio (SNR) of 32.76%, 19.02%, and 8.16% across the M4Raw, in vivo and phantom datasets, respectively. Qualitative assessments, including difference maps, line intensity plots, and effective receptive fields, suggested that NND preserves structural details and edges compared to random noise denoising (RND), indicating potential enhancements in visual quality. This substantial improvement in low-field imaging quality addresses the fundamental challenge of diagnostic confidence in resource-constrained settings. By mitigating the primary technical limitation of these systems, our approach expands the clinical utility of low-field MRI scanners, potentially facilitating broader access to diagnostic imaging across resource-limited healthcare environments globally.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".