AI RESEARCH

Uni-Encoder Meets Multi-Encoders: Representation Before Fusion for Brain Tumor Segmentation with Missing Modalities

arXiv CS.CV

ArXi:2604.22177v1 Announce Type: new Multimodal MRI offers complementary information for brain tumor segmentation, but clinical scans often lack one or modalities, which degrades segmentation performance. In this paper, we propose UniME (Uni-Encoder Meets Multi-Encoders), a two-stage heterogeneous method for brain tumor segmentation with missing modalities that reconciles the trade-offs among fine-grained structure capture, cross-modal complementarity modeling, and exploitation of available modalities.