Res-MoCoDiff: Residual-guided diffusion models for motion artifact correction in brain MRI

2025

Read Res-MoCoDiff: Residual-guided diffusion models for motion artifact correction in brain MRI (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-03498.txt.

Description

Mojtaba Safari and co-authors present a diffusion model that cleans motion blur out of brain scans, in a 2025 paper. When a patient moves during an MRI, the image smears and doctors may need a repeat scan. Their model, Res-MoCoDiff, learns from the leftover error between clean and smeared images. It folds that error into its noise process, so the cleanup run needs only four steps instead of many.

The network uses windowed attention blocks that cope with varied image sizes. Training mixes two error measures to keep images sharp and true. On fake motion made by a realistic simulator and on a real patient data set, the model beat three rivals. It scored best on likeness and error marks at every smear level, and it cleaned a pair of slices in a fraction of a second, far faster than standard diffusion tools.

Catalog record

YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-03498

Browse the full library · About the Tower of Babel