【摘要】 目的:使用知识图谱的方法分析医学影像领域肝脏分割研究的知识基础、研究热点。方法:以Web of Science核心数据集为来源,将"肝脏""医学影像""CT""MRI""PET""超声"等作为主题词,"分割"作为题目关键词进行检索,使用Bicomb 2. 0对引文、关键词等数据进行清洗和频次统计,使用g CLUTO 1. 0构建高频引文、高频关键词双聚类图,使用Excel构建高频引文、高频关键词战略坐标图,使用Ucinet 6构建高频引文、高频关键词共现网络图。结果:共检索到825篇文献,自2007年开始载文量逐年快速增长。高频引文双聚类分析发现5类基础知识,高频关键词双聚类分析发现6个研究热点。结论:肝脏自动分割仍处于初始阶段,虽然形成了一些主题和分割方法,并不断引入新的分割方法,但这些算法各有优势及局限,需要根据具体情况选择、设计和组合形成合适的分割算法。更多还原

【Abstract】 Objective To analyze the knowledge and hotspots in researches on liver segmentation according to their knowledge graphs. Methods The papers on liver segmentation were retrieved with Web of Science core data set as the data source,with liver,medical images,CT,MRI,PET and ultrasound as the subject headings,and with segmentation as the key word in titles. The data,such as citations and key words,were cleaned and their frequencies were calculated using Bicomb 2. 0,the double clustering graphs of high-frequency citations and high-frequency key words were plotted using gCLUTO 1. 0,the strategic coordinate diagrams and the co-occurrence network graphs of high-frequency citations and high-frequency key words were established using Excel and Ucinent 6 respectively. Results Eight hundred and twenty-five papers were retrieved. The number of published papers on liver segmentation has rapidly increased year by year since 2007. Double clustering analysis of high-frequency citations showed 5 kinds of basic knowledge,and double clustering analysis of high-frequency key words displayed 6 hotspots in researches on liver segmentation. Conclusion Automatic liver segmentation is still in its initial stage. Although certain themes and segmentation methods have formed,and new segmentation methods have been introduced. However,such algorithms have their own advantages and disadvantages,it is thus necessary to select,design and integrate the appropriate segmentation algorithms according to the specific conditions.更多还原