骨折影像数据标注与质量控制专家共识
骨折影像数据标注与质量控制专家组
王浩1
张巍2
1.102629 北京,中国食品药品检定研究院医疗器械检定所2.100048 北京,中国人民解放军总医院第四医学中心骨科医学部、国家骨科与运动康复临床医学研究中心
摘要：In recent years,artificial intelligence(AI)has been developed rapidly in the field of fracture auxiliary analysis,with applications covering lesion localization,diagnostic decision-making,and surgical planning.Large-scale,high-quality,and highly consistent labeled data would facilitate the improvement of AI model accuracy and generalization.However,currently there is no specific data annotation standard for fracture imaging.To promote industry development and ensure the standardization and quality of annotations,National Institutes for Food and Drug Control(NIFDC)and the Fourth Medical Center of the People's Liberation Army General Hospital have jointly researched data annotation and quality control methods for fracture imaging.This article focuses on the annotation rules,annotation process,and quality control of fractures,introducing the preliminary consensus reached by the expert group.It aims at promoting the standardization of relevant dataset construction,improve the consistency and quality of data annotation,and provide strong support for the research,development,and testing of related AI products.
关键词：人工智能医疗器械骨折质量控制多数赞同数据标注
分类号：R814.4(放射医学)R816.8(放射医学)
资助基金：国家骨科与运动康复临床医学研究中心创新基金(23-NCRC-CXJJ-ZD-2-4)国家重点研发计划(2023YFC2508800)国家自然科学基金(82302788)
论文发表日期：2026-02-19
在线出版日期：2026-03-09（本平台首次上网日期，不代表文献的发表时间）
页数：5( 120-124 )
英文信息