Abstract
Accurate brain tumor segmentation from multimodal 3D MRI remains challenging due to heterogeneous modality characteristics and spatially varying image quality across modalities. Most existing fusion strategies typically rely on implicit feature correlation modeling and do not explicitly account for voxel-wise reliability differences during integration. We propose a reliability-aware multimodal segmentation frame-work that estimates spatial modality reliability during stage-wise fusion.A modality-specific Mamba-based encoder captures long-range dependencies while preserving modality-specific representations. A light weight gating mechanism predicts voxel-wise reliability scores using complementary cross-modal cues to guide normalized feature aggregation. Combined with hierarchical decoding, the proposed framework reduces modality interference and improves the delineation of challenging tumor subregions. Experiments on two public benchmarks demonstrate the superiority of our method over state-of-the-art methods.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/BrainWorks_011.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/profile?id=~Jinyan_Fan1
BibTex
@InProceedings{WanXin_Reliabilityaware_MICCAISAT2026,
author = { Wang, Xinyue AND Zhou, Yumou AND Fan, Jinyan},
title = { { Reliability-aware Multimodal Fusion for Brain Tumor Segmentation } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17254},
month = {pending},
page = {pending}
}
