科学图像的单样本自适应分割
One-Shot Adaptive Segmentation For Scientific Images
Tejaswi V. Panchagnula · Allison M. Davis · Fengqing Zhu
中文摘要
科学图像分割方法依赖大量标注与任务专属训练,难以适配不同成像模态与实验条件。提出一种免训练的单样本(one-shot)框架,仅利用一张带标注的参考图像即可将视觉基础模型适配到目标任务。该框架融合 DINOv3 表征与背景自适应特征正交化,以抑制与伪影相关的特征方向,再通过余弦相似度定位候选区域,最后由 SAM 完成分割。在红细胞显微成像、结构光照池沸腾以及胸部 X 光片三个数据集上的评估表明:相对最强基线,显微与池沸腾数据集的 mean IoU 分别提升 5.91% 和 78.62%,在胸部 X 光片上达到相当性能。结果证明,单样本参考条件化能够将通用视觉模型有效适配到专业科学分割任务中。
关键要点
- 01问题:科学图像分割依赖大量标注与任务专属训练,跨模态、跨实验条件适应性差
- 02方法:结合 DINOv3 表征与背景自适应特征正交化抑制伪影方向,以余弦相似度定位候选区域并由 SAM 完成分割,全程免训练
- 03结果:在红细胞显微与池沸腾数据集上 mean IoU 相对最强基线分别提升 5.91% 和 78.62%,胸部 X 光片性能相当
- 04局限:仅基于单张参考图像,效果受参考样本质量与代表性影响;不同成像模态间增益差异显著
解读
尚无解读。
原始英文摘要
arXiv:2610.10306v1 Announce Type: new Abstract: Scientific image segmentation methods rely on extensive annotation and task-specific training, limiting adaptation across imaging modalities and experimental conditions. We present a training-free, one-shot framework that specializes vision foundation models using a single annotated reference image. The framework combines DINOv3 representations with background-adaptive feature orthogonalization to suppress artifact-related feature directions, after which cosine similarity localizes candidate regions for SAM segmentation. We evaluate the framework on red-blood-cell microscopy, structured-illumination pool boiling, and chest radiography. Relative to the strongest baseline, the proposed method improves mean IoU by 5.91% and 78.62% on the microscopy and pool-boiling datasets, respectively, while achieving comparable performance on chest radiographs. These results demonstrate that one-shot reference conditioning can adapt general-purpose vision models to specialized scientific segmentation tasks.