发表机构
Department of Electrical Engineering and Computer Science, UC Berkeley; Voio; Department of Radiology and Biomedical Imaging, UC San Francisco; Computational Precision Health, UC Berkeley and UC San Francisco（加州大学伯克利分校电气工程与计算机科学系; Voio; 加州大学旧金山分校放射学与生物医学成像系; 加州大学伯克利分校和加州大学旧金山分校计算精准健康中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究基于语言弱监督训练CT扫描文本条件分割模型，通过结合体素级与切片级监督，从扫描-报告对提取信息，微调通用2D图像分割模型SAM3，在ReXGroundingCT数据集上提升了分割骰子分数。
AI 中文摘要
我们介绍了一种用于训练CT扫描的文本条件分割模型的方法，该方法将体素级监督与来自报告的粗略但可扩展的切片级监督相结合。我们从大量扫描-报告对数据库中提取带有发现描述及出现这些发现的切片索引。然后，我们使用来自强标记数据的标准分割损失和来自提取的弱监督的切片级分类损失，对通用2D图像分割模型SAM3进行微调。我们在ReXGroundingCT数据集上的结果表明，这种策略提高了分割骰子分数：从1000个完全标记体积时的8%相对增益到250个完全标记体积时的22%。
英文摘要
We introduce a method for training a text-conditioned segmentation model for CT scans, which combines voxel-level supervision with coarse but scalable slice-level supervision from reports. We extract, from a large database of scan-report pairs, descriptions of findings with indices of slices where those findings occur. We then finetune a general-purpose 2D image segmentation model, SAM3, with standard segmentation losses from strongly labeled data and with a slice-level classification loss from the extracted weak supervision. Our results on the ReXGroundingCT dataset illustrate that this strategy improves the segmentation dice score: from an 8% relative gain when there are 1000 fully labeled volumes to 22% when there are 250 fully labeled volumes.
