CXR-PRO: MIMIC-CXR with Prior References Omitted

CXR-PRO is an adaptation of the MIMIC-CXR dataset that omits references to prior radiology reports. Consisting of 374,139 free-text radiology reports and associated chest radiographs, CXR-PRO addresses the issue of hallucinated references to priors produced by radiology report generation models. By removing nearly all prior references in MIMIC-CXR, CXR-PRO, when used as training data for report generation models, is capable of broadly improving the factual consistency and accuracy of generated reports. More generally, this dataset aims to support a wide body of research in medical image analysis and natural language processing. MIMIC-CXR is a de-identified dataset, so no protected health information (PHI) is included.CXR-PRO是MIMIC-CXR数据集的适配版本，其移除了所有既往放射报告的相关引用。该数据集包含374139份自由文本放射报告及配套胸部X光影像，解决了放射报告生成模型所产生的对既往报告的幻觉引用问题。通过移除MIMIC-CXR中几乎所有的既往引用内容，当将CXR-PRO用作报告生成模型的训练数据时，可全面提升生成报告的事实一致性与准确性。更广泛而言，该数据集旨在为医学图像分析与自然语言处理领域的大量研究提供支撑。MIMIC-CXR是去标识化数据集，因此未包含任何受保护健康信息（Protected Health Information, PHI）。

搜集汇总

背景与挑战

背景概述

CXR-PRO是基于MIMIC-CXR数据集改编的版本，主要特点是通过自动方法去除了放射学报告中对先前报告的引用，以解决报告生成模型中幻觉引用的问题。数据集包含374,139份自由文本放射学报告和相关胸片图像，分为训练集和专家编辑的测试集，旨在提高放射学报告生成模型的事实一致性和准确性，适用于医学图像分析和自然语言处理研究。