摘要
目的应用CiteSpace可视化软件系统分析近十年超声医学领域人工智能的研究热点和趋势。方法以中国知网、万方、PubMed和Web of Science为来源数据库,检索2014~2024年超声医学领域人工智能相关文献,应用CiteSpace可视化软件对纳入文献的发文量、作者、发文机构、国家、关键词等绘制图谱,并进行计量分析。结果最终纳入3081篇文献,其中中文文献1418篇,外文文献1663篇。近十年超声医学领域人工智能的发文量整体呈逐年上升趋势,尤其在2018年后增长迅速,相关研究进入快速发展阶段并趋于成熟。中文文献和外文文献作者的合作网络密度值分别为0.0076和0.0073,机构的合作网络密度值分别为0.0029和0.0179,表明作者及机构之间缺乏密切合作。发文机构分析显示文献发表集中在高等教育学校。发文量前5名的国家分别为中国(694篇)、美国(396篇)、韩国(75篇)、加拿大(74篇)、印度(53篇),虽然发文国家之间存在交流合作,但国家的合作网络密度值为0.0141,表明合作密切程度较低。关键词分析显示,中文文献和外文文献热门关键词有一定相似性,集中在深度学习、图像分割、乳腺癌、卷积神经网络、图像分类、甲状腺结节等方面。结论近十年国内外超声医学领域人工智能的相关研究取得了显著进展,研究热点主要集中在深度学习、图像分割和智能辅助识别与诊断等方面,研究趋势将聚焦于基于超声造影的人工智能模型。
Objective To systematically analyze the research hotspots and trends of AI in ultrasound medicine over the past decade by CiteSpace visualization software.Methods Literature related to AI in ultrasound medicine published between 2014 and 2024 was retrieved from the CNKI,Wanfang,PubMed,and Web of Science databases.CiteSpace visualization software was used to visualize and quantitatively analyze the graphs of publication volume,authors,institutions,countries,and keywords.Results A total of 3081 articles were included,comprising 1418 Chinese and 1663 non-Chinese publications.Over the past decade,the number of publications regarding AI in ultrasound medicine showed an overall increasing trend,with rapid growth after 2018,indicating a phase of accelerated development and maturation.The collaboration network densities among authors were 0.0076 for Chinese literature and 0.0073 for non-Chinese literature,while the institutional collaboration network densities were 0.0029 and 0.0179,respectively,reflecting insufficient collaboration among authors and institutions.Analysis of institutions revealed a concentration in higher education institutions.The top five countries by publication volumn were China(694 articles),the USA(396 articles),South Korea(75 articles),Canada(74 articles),and India(53 articles).Although international collaboration among the issuing countries existed,the collaboration network density was 0.0141,indicating a relatively low degree of cooperation.Keyword analysis indicated that similarities in popular topics between Chinese and non-Chinese literature,focusing on deep learning,image segmentation,breast cancer,convolutional neural network,image classification,and thyroid nodules.Conclusion Significant progress has been made in AI research in ultrasound medicine over the past decade.Research hotspots primarily include deep learning,image segmentation,intelligent-assisted identification and diagnosis.Future trends are expected to focus on AI models based on contrast-enhanced ultrasonography.
作者
刘文丽
李宁
王艺桦
张海鸥
章若瑶
张树华
LIU Wenli;LI Ning;WANG Yihua;ZHANG Haiou;ZHANG Ruoyao;ZHANG Shuhua(Department of Ultrasound,North China University of Science and Technology Affiliated Hospital,Tangshan 063000,China)
出处
《临床超声医学杂志》
2025年第10期833-840,共8页
Journal of Clinical Ultrasound in Medicine
基金
《超声诊断学》精品教学案例(库)建设基金项目(KCJPZ2023030)。
