Uncertainty-aware feature mapping and adaptive inference for brain tumor segmentation with missing contrast-enhanced T1-weighted imaging

Weidong Liu, Zhuyin Zhang, Fangfang Deng, Yeyu Xiao · 2026

Read Uncertainty-aware feature mapping and adaptive inference for brain tumor segmentation with missing contrast-enhanced T1-weighted imaging by Weidong Liu, Zhuyin Zhang, Fangfang Deng, Yeyu Xiao (2026) 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/PMC13319089.txt.

Description

Weidong Liu and colleagues present a 2026 framework for segmenting brain tumours on MRI when the contrast-enhanced scan is missing. That sequence is often unavailable because contrast agents are unsafe for some patients or scans fail. Performance of standard tools drops sharply without it. Their system, named UAF-AIMM, tackles the gap in three parts.

One module rebuilds the missing sequence's features from the scans that exist. A second weighs each spot by estimated reliability before fusing features. A third recalibrates the model for each new case at test time. On the BraTS 2021 benchmark, the system reached a Dice score of 89.20 percent in their testing. Gains were largest in the enhancing tumour region, the part most dependent on contrast. The approach outperformed four named baselines.

Catalog record

CREATORWeidong Liu, Zhuyin Zhang, Fangfang Deng, Yeyu Xiao
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE66059
LICENSEhttps://creativecommons.org/licenses/by/4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/article/MED/PMC13319089

Browse the full library · About the Tower of Babel