arXiv 2502.165852025-02-23cs.CV

解剖学接地预训练用于医学短语接地

Anatomical grounding pre-training for medical phrase grounding

University of Queensland（昆士兰大学）；Australian e-Health Research Centre, CSIRO Health and Biosecurity（澳大利亚联邦科学与工业研究组织健康与生物安全部澳大利亚电子健康研究中心）

Wenjun Zhang, Shakes Chandra, Aaron Nicolson

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出解剖学接地预训练，利用Chest ImaGenome数据集对齐解剖术语与图像区域，在MS-CXR上显著提升医学短语接地性能，微调模型mIoU达61.2，超越现有最先进方法。

摘要：

医学短语接地（MPG）将医学报告中描述的放射学发现映射到医学图像中的特定区域。阻碍MPG进展的主要障碍是可用于训练和验证的标注数据稀缺。我们提出解剖学接地作为一项领域内预训练任务，将解剖学术语与医学图像中的相应区域对齐，利用大规模数据集如Chest ImaGenome。我们在MS-CXR上的实证评估表明，解剖学接地预训练在零样本学习和微调设置中均显著提高了性能，超越了最先进的MPG模型。我们的微调模型在MS-CXR上取得了最先进的性能，mIoU达到61.2，证明了解剖学接地预训练对MPG的有效性。

英文摘要：

Medical Phrase Grounding (MPG) maps radiological findings described in medical reports to specific regions in medical images. The primary obstacle hindering progress in MPG is the scarcity of annotated data available for training and validation. We propose anatomical grounding as an in-domain pre-training task that aligns anatomical terms with corresponding regions in medical images, leveraging large-scale datasets such as Chest ImaGenome. Our empirical evaluation on MS-CXR demonstrates that anatomical grounding pre-training significantly improves performance in both a zero-shot learning and fine-tuning setting, outperforming state-of-the-art MPG models. Our fine-tuned model achieved state-of-the-art performance on MS-CXR with an mIoU of 61.2, demonstrating the effectiveness of anatomical grounding pre-training for MPG.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。