融合先验知识和信息化技术的望诊客观化研究现状与思路探析  

The Research Status and Thinking of Objectification of Diagnosis by Integrating Prior Knowledge and Information Technology

作　　者：孙晨格 李玉坤 赵志伟 汪麟双 崔梁瑜 尹彤 王丽颖[2] 韩学杰[2] 刘大胜[2] SUN Chenge;LI Yukun;ZHAO Zhiwei;WANG Linshuang;CUI Liangyu;YIN Tong;WANG Liying;HAN Xuejie;LIU Dasheng(Shaanxi University of Chinese Medicine,Xianyang 712000,China;Institute of Clinical Basic Medicine of Traditional Chinese Medicine,Chinese Academy of Traditional Chinese medicine,Beijing 100700,China)

机构地区：[1]陕西中医药大学,咸阳712000 [2]中国中医科学院中医临床基础医学研究所,北京100700

出　　处：《世界科学技术-中医药现代化》2024年第6期1654-1659,共6页Modernization of Traditional Chinese Medicine and Materia Medica-World Science and Technology

基　　金：国家自然科学基金委员会青年科学基金项目(82205320):融合图像识别技术与先验知识的冠心病患者大鱼际望诊深度学习模型研究,负责人:刘大胜;第七批全国名老中医药专家学术经验继承项目(No.Z0814):韩学杰全国老中医药专家传承工作室,负责人:韩学杰;中国中医科学院优秀青年科技人才培养专项(ZZ16-YQ-035):基于人工智能的冠心病患者大鱼际望诊深度学习模型临床验证及机制研究,负责人:刘大胜。

摘　　要：望诊客观化主要采用图像处理、计算机视觉和机器学习等技术方法,以解决望诊主观性强和难以量化的问题。而先验知识的获取与处理是进行望诊客观化的关键环节,也是对客观化研究中将主观判断和宏观表现进行量化展示的重要阐述,但目前仍缺乏对先验知识的深入总结及参数化处理,本文在分析望诊客观化研究现状的基础上,运用数据挖掘技术将中医望诊经验进行整理归纳,并通过自然语言处理、表示学习法将望诊观察的信息转化为可量化的数字特征,同时深度学习的应用可实现对望诊图像的自动化诊断和分析,以提高准确性和效率,促进中医现代化的进程。In order to solve the problem of strong subjectivity and difficulty in quantification,clinical objectification mainly adopts the techniques of image processing,computer vision and machine learning.The acquisition and processing of prior knowledge is a key link in the objectification of inspection,as well as an important elaboration of the quantification of subjective judgment and macro performance in objectification research.However,there is still a lack of in-depth summary and parametric processing of prior knowledge.Based on the analysis of the current research status of objectification of inspection,this paper uses data mining technology to summarize the experience of TCM inspection.Moreover,the observation information can be transformed into quantifiable digital features through natural language processing and representation learning.Meanwhile,the application of deep learning can realize automatic diagnosis and analysis of observation images to improve accuracy and efficiency,and promote the process of TCM modernization.

关 键 词：望诊 望诊客观化 先验知识 深度学习 参数化 

分 类 号：R241.2[医药卫生—中医诊断学]