用经典滤波器与冻结特征融合，仅凭少量标注实现高精度显微图像分割。
Exemplar: Classical Priors Complement Frozen Features for Few-Shot Microscopy Segmentation at Native Resolution
- 融合冻结的自监督特征与固定经典滤波响应，轻量级统一头结构
- 单个配置在11个数据集上达0.782的分割性能，优于单独使用任一模块
- 仅需1个标注掩码即表现优异，适合资源受限的医学图像分析
为新生物医学数据集进行分割通常需要大量标注训练专属模型，或在推理时调整基础模型。我们提出 Exemplar，一种少样本分割器，将冻结的 DINOv3 主干与一组固定的原生分辨率经典滤波响应融合于一个轻量级头部中，仅需支持掩码即可训练。在少掩码、原生分辨率场景下，经典先验与冻结自监督特征互补：融合后单一固定配置可覆盖11个生物医学成像数据集。同一头部中，经典滤波器单独表现得分为0.693（按前景交并比或中心线Dice），冻结特征为0.672；前者在7个数据集领先，后者在其余4个领先，融合后达到0.782。相比5种前向传播少样本方法，Exemplar在55次方法-数据集对比中领先54次（52次经Holm校正显著）。仅用1个标注掩码即达0.703得分，优于在同一掩码上从零训练的nnU-Net（0.682）。在8个掩码时，nnU-Net在面板均值上反超，主要体现在中心线一致性，但训练耗时为Exemplar的16至77倍。
原文摘要 · Abstract (English)
Segmenting a new biomedical dataset usually means a domain-specific model trained on substantial annotation, or a foundation model steered at inference time. We present Exemplar, a few-shot segmenter that fuses a frozen DINOv3 backbone with a fixed bank of classical native-resolution filter responses in one lightweight head, fitted from the support masks alone. In the few-mask, native-resolution regime, classical priors and frozen self-supervised features are complementary: fused in one head, a single fixed configuration spans eleven biomedical imaging datasets. Under the same head, the classical bank alone reaches 0.693 on the eleven-dataset panel, scored by foreground intersection-over-union or centreline Dice, and the frozen features alone 0.672; the bank leads on seven of the eleven and the features on the rest, and fused they reach 0.782. Against five forward-pass few-shot methods, Exemplar leads in 54 of 55 method-dataset comparisons, 52 of them significant after Holm correction. From a single annotated mask it reaches 0.703 on the same panel, against 0.682 for a from-scratch nnU-Net trained on that same mask. At eight masks nnU-Net overtakes it on the panel mean, chiefly on centreline agreement, but takes 16-77x longer to fit.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。