RadFusion
收藏资源简介:
RadFusion是一个大规模的多模态肺栓塞数据库,包含1837名患者的CT影像研究及其对应的电子健康记录(EHR)数据,旨在通过结合3D医学影像数据和患者EHR数据,推动未来多模态融合策略的研究。该数据集通过分层随机抽样从原始的108,991项研究中精心筛选而出,确保了数据的高质量和代表性。数据集的应用领域包括构建更好的临床决策模型以检测肺栓塞,以及开发结合CT扫描和患者EHR的多模态融合模型,这些在医学AI领域相对未被充分探索,但对实际临床环境中的医学影像解读至关重要。
RadFusion is a large-scale multimodal pulmonary embolism database containing CT imaging studies and corresponding electronic health record (EHR) data from 1837 patients. It aims to advance research on future multimodal fusion strategies by integrating 3D medical imaging data and patient EHR data. The dataset was meticulously screened from the original 108,991 studies via stratified random sampling, ensuring high data quality and representativeness. Its application areas include constructing more robust clinical decision-making models for pulmonary embolism detection, as well as developing multimodal fusion models that combine CT scans and patient EHR data. These areas are relatively under-explored in the field of medical AI, yet are critically important for medical image interpretation in real-world clinical scenarios.




