中医手诊多模态信息客观化设备的研发  

Development of objective multimodal information acquisition equipment for traditional Chinese medicine hand diagnosis

作　　者：周炜翔 徐望明[1] 张晓雨 张晓维 曹祖懋 李筱 关之玥 王爽秋 谭雪琪 张宇格 陈梦珍 但愿 商洪才 徐雪[5] ZHOU Weixiang;XU Wangming;ZHANG Xiaoyu;ZHANG Xiaowei;CAO Zumao;LI Xiao;GUAN Zhiyue;WANG Shuangqiu;TAN Xueqi;ZHANG Yuge;CHEN Mengzhen;DAN Yuan;SHANG Hongcai;XU Xue(School of Information Science and Engineering,Wuhan University of Science and Technology,Wuhan 430081,China;Institute of Basic Research in Clinical Medicine,China Academy of Chinese Medical Sciences,Beijing 100700,China;College of Integrated Traditional Chinese and Western Medicine,Hunan University of Chinese Medicine,Changsha 410208,China;Dongzhimen Hospital,Beijing University of Chinese Medicine,Beijing 100700,China;Medical Department,Wuhan University of Science and Technology,Wuhan 430070,China;Dongfang Hospital,Beijing University of Chinese Medicine,Beijing 100071,China)

机构地区：[1]武汉科技大学信息科学与工程学院,武汉430081 [2]中国中医科学院中医临床基础医学研究所,北京100700 [3]湖南中医药大学中西医结合学院,长沙410208 [4]北京中医药大学东直门医院,北京100700 [5]武汉科技大学医学部,武汉430070 [6]北京中医药大学东方医院,北京100071

出　　处：《中华中医药杂志》2025年第1期83-86,共4页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家重点研发计划(No.2022YFC3502300,No.2022YFC3502302)。

摘　　要：目的:旨在开发一种手诊多模态智能采集设备,建立标准化的手诊多模态信息数据集,并研发图像预处理算法以提高数据质量。方法:融合中医先验知识与先进传感器技术,开发硬件设备;利用深度学习算法规范化采集流程;收集包含800名受试者的手诊数据集;研发预处理算法,提升数据质量。结果:客观指标评估显示,分割算法的边缘保持指数为1.4193,可见光增强算法的对比度增强指数为1.3445,静脉图像增强算法的局部对比度增强指数为1.5196,结构相似性指数为0.9875,均达到标准。结论:研发的设备规范了采集过程,能够采集高质量多模态手诊信息。手部数据的预处理算法显著提升了图像质量,为后续冠心病诊断算法的开发提供了设备和数据基础。Objective:To develop a multimodal intelligent acquisition device for hand diagnosis,establish a standardized dataset of multimodal hand diagnostic information,and develop image preprocessing algorithms to improve data quality.Methods:By integrating traditional Chinese medicine prior knowledge with advanced sensor technology,hardware equipment was developed.Deep learning algorithms were employed to standardize the acquisition process.A hand diagnosis dataset containing 800 subjects was collected.Preprocessing algorithms were developed to enhance image quality.Results:Objective metric evaluations showed that the segmentation algorithm's EPI was 1.4193,the CEI was 1.3445,the LCI was 1.5196,and the structural similarity index was 0.9875,all reaching objective standards.Conclusion:The developed equipment standardized the acquisition process and can collect high-quality multimodal hand diagnosis information.The preprocessing algorithms for hand data significantly improved image quality,providing equipment and data foundations for the subsequent development of coronary heart disease diagnosis algorithms.

关 键 词：手诊 先进传感器 四诊合参 深度学习 多模态 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] R241[自动化与计算机技术—计算机科学与技术]