时间标签：数据日期 2026-10-31 · 生成时间 2026-11-01 05:30（Asia/Shanghai）｜ 检索窗口 PubMed edat 2026/10/28–2026/10/31 ｜ 情报发生日期 2026-10-31
研究问题
XX, XXXX 2020 1 DuMeta++: Spatiotemporal Dual Meta-Learning for Generalizable Few-Shot Brain Tissue Segmentation Across Diverse Ages Yongheng Sun, Jun Shu, Jianhua Ma, and Fan Wang Abstract— Accurate segmentation of brain tissues from MRI scans is critical for neuroscience and clinical ap- plications, but achieving consistent performance across the human lifespan remains challenging due to dynamic, age-related changes in brain appearance and morphology.；Yet, a major challenge stems from the highly folded and spatiotemporally evolving brain morphologies across different ages [2].；DuMeta++ Paradigm We propose a dual meta-learning (DuMeta++) framework, illustrated in Fig. 2, to address the challenge of brain tis- sue segmentation across the lifespan.
方法
In contrast, deep learning has become the de facto standard: U-Net–style 2D [12] and 3D [3] networks and their variants (with attention [13], dense connections [14], and multi-modal fusion [15]) report strong Dice scores across adult cohorts, and self-configuring frameworks such as nnU- Net [16] further simplify model design.；In the meta-test phase, iBEAT is similarly applied, followed by expert manual refinements to produce gold-standard annotations.；The datasets used in this study are briefly summarized: • OASIS3 [34]: This public dataset comprises 1,639 T1w MRIs from 992 individuals, each with 1 to 5 scans captured within a 5-year window.；Participants include 6 IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL.
核心结论
This work significantly extends our prior conference pub- lication [9] through four key advancements: a memory-bank- based class-aware regularization mechanism enhancing lon- gitudinal consistency, formal convergence guarantees for the dual meta-learning pipeline, expanded ablation studies validat- ing design choices, and comprehensive benchmarking against state-of-the-art methods across diverse datasets.；Unlike previous self-supervised learning methods [7], [8] relying on image or patch-level contrastive objectives, we em- ploy a weak-supervised contrastive framework (i.e., according to pseudo labels) that aligns features of each tissue class directly—thereby simplifying the link to the final segmentation goal.；Suppose we have tissue label maps M (or pseudo-labels), which are downsam- pled to match each Fk, producing Mk. ‖ 结论：In conclusion, we have presented DuMeta++, a unified dual meta-learning framework for generalizable, few-shot brain tissue segmentation across varying age groups.
局限
XX, XXXX 2020 1 DuMeta++: Spatiotemporal Dual Meta-Learning for Generalizable Few-Shot Brain Tissue Segmentation Across Diverse Ages Yongheng Sun, Jun Shu, Jianhua Ma, and Fan Wang Abstract— Accurate segmentation of brain tissues from MRI scans is critical for neuroscience and clinical ap- plications, but achieving consistent performance across the human lifespan remains challenging due to dynamic, age-related changes in brain appearance and morphology.；Experiments on diverse datasets (iSeg- 2019, IBIS, OASIS, ADNI) under few-shot settings demon- strate that DuMeta++ outperforms existing methods in cross-age generalization.；Index Terms— Brain Tissue Segmentation, Lifespan, Meta Learning, Cro
对本项目的价值
★☆☆☆☆ 脑老化是最大未满足需求，也是终点最难量化的领域；biomarker 与认知读数选择最敏感。队列：ADNI。
全文来源
出版商 OA PDF（正文 67,454 字符，已过 DOI 逐字 / 标题指纹归属校验）
原文题录
- 英文标题：DuMeta++: Spatiotemporal Dual Meta-Learning for Generalizable Few-Shot Brain Tissue Segmentation Across Diverse Ages.
- 期刊与标识：IEEE Trans Med Imaging · PMID 42955762 · DOI 10.1109/tmi.2026.3748792
A 级证据分级：A 级（同行评议人体研究）
本文属人体研究证据，但需注意：观察性或小样本干预的结论不等于已证实临床获益；效应量与长期安全性仍需更大规模、更长随访的试验验证。
关于本文的文献来源：本站依据 IMA 知识库「五刊监控·深度文献卡片」撰写，依据已获取的期刊全文（方法、样本量与结论已回到原文核对）；未能获取全文的同批条目另行标注「摘要级」，全程未做任何数值推算补全。