基于多维图像分析技术的腺病毒感染中医舌象特征分析  

Analysis of TCM Tongue Image Characteristics of Adenovirus Infection Based on Multi-dimensional Image Analysis Technology

作　　者：荣文雅 王仲霞 何婷婷 景婧 王丽苹 孙永强 朱云 许文涛 桑秀秀 田淼 刘力政 冷源铭 杨伟[4] 王睿林 Rong Wenya;Wang Zhongxia;He Tingting;Jing Jing;Wang Liping;Sun Yongqiang;Zhu Yun;Xu Wentao;Sang Xiuxiu;Tian Miao;Liu Lizheng;Leng Yuanming;Yang Wei;Wang Ruilin(School of Chinese Medicine,Southern Medical University,Guangzhou 510515,China;The Fifth Medical Center of PLA General Hospital,Beijing 100039,China;Institute of Engineering and Applied Technology,Fudan University,Shanghai 200433,China;Institute of Basic Research in Clinical Medicine,China Academyof ChineseMedical Sciences,Beijing100700,China)

机构地区：[1]南方医科大学中医药学院,广州510515 [2]解放军总医院第五医学中心,北京100039 [3]复旦大学工程与应用技术研究院,上海200433 [4]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》2022年第11期4580-4588,共9页Modernization of Traditional Chinese Medicine and Materia Medica-World Science and Technology

基　　金：解放军总后勤部卫生部科研基金(17BJZ53):基于代谢组学分析的酒精性肝病体质差异研究,负责人:王睿林;国家自然科学基金面上项目(81673806):基于呼出气代谢组学的肝衰竭“肝臭”中医嗅诊辨证客观化研究,负责人:王睿林;军委后勤保障部卫生局中医药服务能力培育与提升专项计划面上项目(2021ZY046):中药复方“克冠1号”阻断新型冠状病毒肺炎重症化的作用机制研究负责人:王仲霞。

摘　　要：目的 利用多维图像分析技术对腺病毒感染患者的舌象特征进行处理分析,将舌诊图像转换为客观化数据,为中医诊断及进一步辨证论治提供科学支撑。方法 收集85例腺病毒感染患者以及30例正常人的临床资料,采用舌诊图像数字分析诊断系统对患者的舌象进行采集,并利用多维图像分析技术从齿痕有无、舌苔颜色、舌苔裂纹占比、舌苔湿度、舌苔纹理(纹理粗糙度,纹理方向度,纹理对比度)5个方面提取舌象特征,使用描述性统计和统计检验方法比较正常人群、腺病毒感染发热患者、密切接触患者、无发热患者的舌象特征差异,并采用随机森林算法对舌象特征做重要性排序,筛选重要特征纳入模糊K-Means聚类分析对腺病毒感染者的舌象特征进行归纳。结果 舌象特征分布显示,正常人群、腺病毒感染发热患者、密切接触患者、无发热患者4类人群中舌苔颜色R/G/B、齿痕的有无、舌苔湿度、纹理粗糙度、纹理方向、纹理对比度的分布具有统计学差异,舌苔颜色R/G均值、纹理粗糙度、纹理对比度可能对4类人群具有较好的区分度。结论 基于观察到的腺病毒感染患者舌象,根据多维图像分析技术提取的舌象特征与临床辨证具有较强的一致性,可为中医舌象辨证提供客观依据。Objective To process and analyze the tongue image characteristics of patients with adenovirus infection by using multi-dimensional image analysis technology, and convert the tongue diagnosis image into objective data, so as to provide scientific support for TCM diagnosis and further syndrome differentiation and treatment. Methods The clinical data of 85 patients with adenovirus infection and 30 normal subjects were collected. The tongue images of patients were collected by tongue diagnosis image digital analysis and diagnosis system, and the multi-dimensional image analysis technology was used to analyze the tooth marks, tongue coating color, proportion of tongue coating cracks, tongue coating humidity and tongue coating texture(texture roughness, texture orientation and texture contrast) Five aspects were used to extract the tongue image characteristics. Descriptive statistics and statistical test methods were used to compare the tongue image charactristics of normal people with fever caused by adenovirus infection, close contact patients and non fever patients. At the same time, the tongue image characteristics were screened by the importance of random forest algorithm and included in fuzzy k-means cluster analysis to summarize the tongue image characteristics of adenovirus infected patients. Results The distribution of tongue image characteristics showed that there were significant differences in tongue coating color R/G/B, tooth mark, tongue coating humidity, texture roughness, texture direction and texture contrast among normal people, patients with fever caused by adenovirus infection, close contact patients and patients without fever Texture contrast may have a good distinction between the four groups. Conclusion Based on the observed tongue images of patients with adenovirus infection, the tongue image features extracted according to multi-dimensional image analysis technology are consistent with clinical syndrome differentiation, which can provide an objective basis for tongue image syndrome differentiat

关 键 词：腺病毒 舌象 多维图像分析技术 舌诊客观化 

分 类 号：R319[医药卫生—基础医学]