Progressive multi-scale tumor-state memory for incomplete multimodal brain tumor segmentation
We propose a progressive multi-scale tumor-state memory network (PMTM) for segmentation from any non-empty subset of T1ce, T1, FLAIR, and T2.
ProofPaper ↗
Key points
- Brain tumor segmentation uses complementary magnetic resonance imaging (MRI) sequences, but some sequences may be unavailable in practice.
- Their absence changes the information available for segmenting different tumor regions.
- Fusion methods can support different input subsets.
- We evaluate PMTM on BraTS2018, BraTS2020, and BraTS2023 GLI across all 15 non-empty modality subsets.
Sources (1)
- [1]Progressive multi-scale tumor-state memory for incomplete multimodal brain tumor segmentationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 02:11 PM
We propose a progressive multi-scale tumor-state memory network (PMTM) for segmentation from any non-empty subset of T1ce, T1, FLAIR, and T2.
Brain tumor segmentation uses complementary magnetic resonance imaging (MRI) sequences, but some sequences may be unavailable in practice.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 19, 2026Sampling of Signals on Multigraphs: From Removable and Uniqueness sets to Spectral Proxies
- Oct 19, 2026Think, Compress, Answer: Reinforcement Learning for Multimodal Grounding via Reasoning Compression
- Oct 18, 2026Where and Why Search-Augmented LLMs Fail When Reasoning over Rapidly Evolving Knowledge
- Oct 17, 2026Measuring Consensus in AI Agent Networks
- Oct 17, 2026Conditional Weight of Evidence for Gradient-Boosted Credit Models: Estimation, Additivity and Its Relation to Adverse Action Reasons
- Oct 16, 2026VocalGrad: Evaluating Acoustic Perception in Audio Language Models