AMM-Diff: Adaptive Multi-Modality Diffusion Network for Missing Modality Imputation

Aghiles Kebaili, Jérôme Lapuyade-Lahorgue, Pierre Vera, Su Ruan · 2025

Read AMM-Diff: Adaptive Multi-Modality Diffusion Network for Missing Modality Imputation by Aghiles Kebaili, Jérôme Lapuyade-Lahorgue, Pierre Vera, Su Ruan (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-2501-12840.txt.

Description

Aghiles Kebaili, Jérôme Lapuyade-Lahorgue, Pierre Vera, and Su Ruan generate missing brain scan types with a diffusion model in a 2025 paper. Tumor segmentation needs several MRI contrasts, but clinics often lack some of them, because protocols are complex or rules restrict scanning. Their model, AMM-Diff, fills in whichever contrasts are missing from whatever is available.

Earlier generators handled only one or two target types. A fusion network first learns a shared representation from all inputs plus their sharp frequency details, taught through a self-supervised task. The diffusion model then rebuilds the missing scans guided by that representation. Tests on the BraTS 2021 brain tumor data set showed the approach works across input combinations.

Catalog record

CREATORAghiles Kebaili, Jérôme Lapuyade-Lahorgue, Pierre Vera, Su Ruan
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-12840

Browse the full library · About the Tower of Babel