【作者】 彭瑜；

【导师】 李瀚旻；

【作者基本信息】 湖北中医药大学 ， 中医内科学， 2014， 博士

【摘要】 目的：构建数字化面色望诊信息系统，用于研究慢性乙型肝炎相关病证与面部五色的变化规律，为提高中医临床诊疗水平提供辅助工具。方法：通过文献研究和专家咨询，根据“五色主五病”中医理论认识和“慢性肝病面容”相关的临床实践，探讨慢性乙型肝炎数字化面色望诊的基本原理和规律，以及面部五色的变化与慢性乙型肝炎相关病证生物学指标的相关性。运用数字化面色望诊技术建立面部望诊五色信息的客观、量化识别方法，实现传统诊法的客观化和定量化，以提高慢性乙型肝炎中医的诊疗水平。通过文献研究、专家咨询和实验研究，探讨数字图像的特点、影响数字图像采集的因素、以及对数字图像进行质量评价和校正的方法。利用湖北省中医院提供的中医面部采集设备，确定数字化面色望诊图像采集方案。应用软件工程方法进行数字化面色望诊信息系统的构建，在明确软件需求的基础上，建立系统功能图、系统流程图、数据库设计、界面设计。系统实现患者信息的录入、面部图像的预处理，以及面部图像颜色分析等功能，为后续对慢性乙型肝炎患者进行面色望诊，提供数字化面色望诊图像处理平台。参照临床研究方法对2011年6月至2012年12月在湖北省中医院肝病科就诊的慢性乙型肝炎患者，根据2000年中华医学会传染病与寄生虫病学分会、肝病学分会联合修订的《病毒性肝炎防治方案》将慢性乙型肝炎患者分为轻度组、中度组与重度组；参照2006年《肝衰竭诊疗指南》将慢性乙型肝炎患者分为轻症组与肝衰竭组；将符合纳入标准的HBeAg阴性的慢性乙型肝炎肝肾精虚兼夹证患者随机分为地五养肝胶囊治疗组、抗病毒治疗组和地五养肝胶囊联合抗病毒治疗组。运用已构建的数字化面色望诊系统对上述三个不同类别的患者进行观察研究。采集面部图像并记录患者基本情况、症状体征以及实验室检查结果，并进一步研究不同肝炎分级患者、轻症慢性乙型肝炎患者与肝衰竭患者、肝肾精虚兼夹证患者治疗前后的面色数据变化规律。结果：构建的面色图像采集方案，能较好的获取患者面色图像，为不同图像之间进行一致性的色彩研究提供了保障，图像质量能满足研究需要。构建的数字化面色望诊系统具有信息录入功能、图像预处理功能以及面部图像色彩分析功能。信息录入功能可以根据研究需要，录入患者的姓名、年龄、性别等基本信息；录入患者ALT、AST、GGT、ALP、TP、ALB、A/G、TBIL、DBIL、IBIl、HBV‐DNA等实验室检查结果；录入患者面色无华、面色晦暗、黄疸黄色鲜明、黄疸黄色晦暗、目黄、口唇、面色青、面色赤、面色黄、面色白、面色黑等症状体征信息；录入采集的患者面部图像信息。图像预处理功能可以完成对采集的图像进行旋转、平移、裁切等操作。而且通过图像预处理功能，根据中医面部对应五脏的理论，可获取具有色彩一致性的面部图像研究区域，能对采集到的不同患者面部图像之间进行一致性的色彩研究。面部图像色彩分析功能对预处理后的患者面部图像，进行RGB色彩分析、HIS色彩分析以及五色指数分析，为数字化面色望诊提供客观、定量的标准化研究数据，为面部望诊提供准确的数字化诊断依据。应用已构建的数字化面色望诊系统对慢性乙肝患者面色进行统计分析。在垂直方向上将双眼下际至鼻孔上际设为D区，在水平方向上将所采集的图像中左眼外侧至左眼内侧设为2区，左眼内侧至右眼内侧设为3区，右眼内侧至右眼外侧为设4区，研究此三个区域（D2、D3、D4）的色彩变化情况。慢乙肝轻度、中度和重度患者面部五色值中，青色指数（D2：11.14±3.33vs.13.98±4.05vs.14.59±4.86；D4：10.42±3.64vs.12.82±3.49vs.13.61±4.14）、黄色指数（D2：13.73±4.43vs.15.59±4.62vs.17.08±5.31；D3：12.47±3.94vs.15.65±4.26vs.16.18±4.70；D4：13.48±4.14vs.15.10±4.28vs.17.93±5.49）、黑色指数（D2：8.70±2.16vs.13.69±4.61vs.18.43±4.20；D3：8.51±3.07vs.10.25±3.38vs.14.74±4.96；D4：8.74±3.34vs.14.44±4.57vs.18.14±5.14）均随病情的加重而明显增加，经统计学处理，差异显著，p＜0.05。肝衰竭患者与轻症慢乙肝患者面部五色值中，青色指数（D2：10.29±4.06vs.15.61±4.14；D3：10.75±4.65vs.15.45±4.94；D4：10.31±4.03vs.15.55±4.79）、黄色指数（D2：11.38±4.86vs.17.93±5.49；D3：11.41±4.43vs.17.57±5.56；D4：11.63±4.47vs.17.23±5.83）、黑色指数（D2：9.44±4.14vs.17.14±5.14；D3：9.76±4.66vs.17.60±5.61；D4：9.38±4.92vs.17.87±5.49）均随病情的加重而明显增加，经统计学处理，差异显著，p＜0.05。地五养肝胶囊治疗组（DWYG）、抗病毒治疗组（NA）和地五养肝胶囊联合抗病毒治疗组（DWYG＋NA）三组之间治疗前的面部五色值无显著性差异，p＞0.05；治疗后，DWYG组治疗前后比较，青色指数（D2：14.55±5.62vs.11.64±3.86；D3：14.41±5.09vs.11.93±3.17；D4：13.71±5.01vs.11.13±3.48）、黄色指数（D2：13.09±5.11vs.11.08±3.34；D3：13.03±5.15vs.11.82±3.32；D4：13.52±5.47vs.11.11±3.62）、黑色指数（D2：11.01±5.39vs.10.30±3.83；D3：11.24±5.41vs.10.71±3.59）均明显下降，p＜0.05，具有显著的统计学差异。NA组治疗前后比较，青色指数（D2：14.94±5.65vs.12.27±3.32；D3：14.43±5.78vs.12.03±3.83；D4：13.74±5.19vs.12.79±3.31）、黄色指数（D2：13.83±5.46vs.11.19±3.95；D3：13.97±5.45vs.11.44±3.67；D4：13.97±5.78vs.11.05±3.72）均下降，p＜0.05，具有显著的统计学差异。DWYG＋NA组治疗前后比较，青色指数（D2：14.17±5.52vs.10.11±3.58；D3：14.83±5.55vs.10.73±3.34；D4：13.30±5.53vs.10.57±3.18）、黄色指数（D2：13.76±5.38vs.10.71±3.47；D3：13.82±5.86vs.10.43±3.92；D4：13.66±5.19vs.10.54±3.85）、黑色指数（D2：11.32±5.71vs.9.32±3.45；D3：11.44±5.85vs.9.66±3.22；D4：11.54±5.17vs.9.89±3.18）均下降，p＜0.05，具有显著的统计学差异。三组间比较DWYG＋NA组青色指数（D2：4.06；D3：4.10；D4：2.73）、黄色指数（D2：3.05；D3：3.39；D4：3.12）、黑色指数（D2：2.00；D3：1.78；D4：1.65）变化最明显，p <0.05，具有显著的统计学差异； DWYG组的青色指数（D2：2.91；D3：2.48；D4：2.58）、黄色指数（D2：2.01；D3：1.21；D4：2.41）、黑色指数（D2：0.71；D3：0.53；D4：0.52）变化较NA组的青色指数（D2：2.67；D3：2.40；D4：0.98）、黄色指数（D2：2.64；D3：2.53；D4：2.92）、黑色指数（D2：‐0.10；D3：0.08；D4：0.61）变化无统计学差异。结论：构建的数字化面色望诊系统可以录入患者基本信息、症状体征信息、实验室检查结果信息及面部图像信息；系统能对录入的面部图像进行有效的预处理，完成图像的平移、旋转、裁切、去噪、提取边缘等操作，不同的图像可进行一致性的色彩研究；系统能对预处理后的图像进行RGB分析、HIS分析和五色分析。构建的数字化面色望诊系统可以弥补传统中医诊断方法的不足，为摆脱主观因素对中医诊断的干扰起到有效的帮助，为临床面部望诊提供客观量化的数据，具有一定的应用价值。采用构建的数字化面色望诊系统对慢性乙型肝炎患者的面色进行客观量化研究，其结果可反映肝肾脾之精气衰减的演变规律，对临床诊疗、辅助辨证论治、疗效评价具有一定参考价值，有助于提高中医药的服务能力和水平。更多还原

【Abstract】 Objective:To constructe the digital inspection information system, in order to study the change rules of chronic HBV‐related diseases and provide auxiliary tools to improve the levels of clinical diagnosis and treatment of traditional Chinese medicine (TCM).Methods:Based on the TCM theory of "five colours main five diseases" and the clinical practice for chronic liver disease face, we used the TCM facial acquisition equipment to explore the fundamentals and rules of digital facial inspection of chronic hepatitis B (CHB), and investigate the relevance of changes in facial five colors to the biological indicators of CHB. Moreover, we used facial inspection technology to construct the objective and quantitative identification methods of facial digital information, in order to establish the objective and quantitative diagnosis and treatment methods of CHB in TCM, and improve the level of diagnosis and treatment of CHB in TCM.Through literature study, expert consultation and experimental research, we discussed the characteristics of digital image, the influencing factors on digital image acquisition, the evaluation methods of digital image quality and the methods of correction in digital image. And we also determined the digital facial inspection image acquisition scheme and tested it through using the TCM facial acquisition equipment provided by Hubei provincial hospital of TCM.Software engineering methods were applied to constructing the digital inspection information system. On the basis of the definite software requirements, system function diagram, system flow chart, database design and interfacial design were established. The established system can realize the functions, such as the input of patient information, the preprocessing of facial images and the analysis of facial images, and provide the inspection digital image processing platform for the following facial inspection with CHB patients.Reference to clinical research methods and based on “the prevention and treatment guideline for viral hepatitis” revised by Chinese Society of Infectious Diseases and Parasitic Diseases in association with Chinese Society of Hepatology, Chinese Medical Association in2000, the CHB patients from2011.6to2012.12in Hubei provincial hospital of TCM were divided into three groups: mild group, moderate group and severe group. Based on “the Diagnostic and treatment guidelines for liver failure” in2006, the above CHB patients were divided into the patients with liver failure and without liver failure. In addition, HBeAg negative CHB patients were randomly divided into three groups: Diwu yanggan Capusle (DWYG) treatment group, antiviral treatment group and DWYG combined antiviral treatment group. Through the observation for the above groups with the three different categories, we acquired the facial images and recorded basic information, symptoms and signs, and laboratory results of the above patients. Then we applied the established digital inspection system to studying the facial change rules of different grading and staging of CHB patients, the patients with liver failure and without liver failure, and the patients with liver and kidney essence deficiency and complexion before and after treatment. Results:The established facial image acquisition scheme could better acquire the facial images of patients, which provided the security for the research of color consistency between different images, and obtained the image quality that satisfied the need of the study.The established digital inspection system has information input function, image preprocessing function and facial image color analysis function.According to the research need, the information input function could record the patient’s name, age, gender, and other basic information; the laboratory results of ALT, AST, GGT and ALP, TP, ALB, A/G, TBIL, DBIL, IBIl and HBV‐DNA in patients; the patients with entry complexion, dim complexion, jaundice with bright yellow, jaundice, dark yellow, yellow, oral, complexion and green, face red, pale yellow, white face, complexion black, the complains, symptoms and signs of patients; and the acquired facial image information of patients.The image preprocessing function could complete the operations, such as rotation, translation, cutting, correction of the acquired image. Through the image preprocessing function, according to the TCM theory of “faces corresponding to the five organs”, we can obtain the consistent facial image study area, and can make the following consistent color research with collected facial images for different patients.The facial image color analysis function could make RGB color analysis, HIS color analysis and the the five colors analysis with facial image of patients after pretreatment, which could obtain the objective, quantitative and standardization of digital inspection data, and provide the accurate digital facial inspection diagnosis basis.We applied the established digital inspection system for the analysis of the facial information with CHB patients. On the facial five‐color values of CHB patients in mild, moderate and severe groups, the cyan index（D2：11.14±3.33vs.13.98±4.05vs.14.59±4.86；D4：10.42±3.64vs.12.82±3.49vs.13.61±4.14）, the yellow index （D2：13.73±4.43vs.15.59±4.62vs.17.08±5.31；D3：12.47±3.94vs.15.65±4.26vs.16.18±4.70；D4：13.48±4.14vs.15.10±4.28vs.17.93±5.49）, the black index （D2：8.70±2.16vs.13.69±4.61vs.18.43±4.20；D3：8.51±3.07vs.10.25±3.38vs.14.74±4.96；D4：8.74±3.34vs.14.44±4.57vs.18.14±5.14） significantly increased with the severity of the disease, which showed statistical difference, p<0.05.In the facial five‐color values of CHB patients with liver failure and without liver failure, the cyan indexes（D2：10.29±4.06vs.15.61±4.14；D3：10.75±4.65vs.15.45±4.94；D4：10.31±4.03vs.15.55±4.79）, the yellow indexes（D2：11.38±4.86vs.17.93±5.49；D3：11.41±4.43vs.17.57±5.56；D4：11.63±4.47vs.17.23±5.83）, the black indexes（D2：9.44±4.14vs.17.14±5.14；D3：9.76±4.66vs.17.60±5.61；D4：9.38±4.92vs.17.87±5.49）significantly increased with the severity of the disease, which showed statistical difference, p<0.05.There was no significant difference in the facial five‐color values between DWYG treatment group, NA treatment group with DWYG＋NA treatment group, P＞0.05. In compared to the facial five‐color values of before DWYG treatment, the cyan indexes（D2：14.55±5.62vs.11.64±3.86；D3：14.41±5.09vs.11.93±3.17；D4：13.71±5.01vs.11.13±3.48）, the yellow indexes（D2：13.09±5.11vs.11.08±3.34；D3：13.03±5.15vs.11.82±3.32；D4：13.52±5.47vs.11.11±3.62）, the black indexes（D2：11.01±5.39vs.10.30±3.83；D3：11.24±5.41vs.10.71±3.59）obviously decreased, which showed statistical difference, p<0.05.In compared to the facial five‐color of before NS treatment, the cyan indexes（D2：14.94±5.65vs.12.27±3.32；D3：14.43±5.78vs.12.03±3.83；D4：13.74±5.19vs.12.79±3.31）, the yellow indexes（D2：13.83±5.46vs. 11.19±3.95；D3：13.97±5.45vs.11.44±3.67；D4：13.97±5.78vs.11.05±3.72）obviously decreased, which showed statistical difference, p<0.05.In compared to the facial five‐color of before DWYG+NS treatment, the cyan indexes（D2：14.17±5.52vs.10.11±3.58；D3：14.83±5.55vs.10.73±3.34；D4：13.30±5.53vs.10.57±3.18）, the yellow indexes（D2：13.76±5.38vs.10.71±3.47；D3：13.82±5.86vs.10.43±3.92；D4：13.66±5.19vs.10.54±3.85）, the black indexes（D2：11.32±5.71vs.9.32±3.45；D3：11.44±5.85vs.9.66±3.22；D4：11.54±5.17vs.9.89±3.18）obviously decreased, which showed statistical difference, p<0.05.The availed data were compared among three groups. DWYG＋NA group: the cyan indexes（D2：4.06；D3：4.10；D4：2.73）, the yellow indexes（D2：3.05；D3：3.39；D4：3.12）, the black indexes（D2：2.00；D3：1.78；D4：1.65）obviously changed, which showed statistical difference, p <0.05. DWYG group: the cyan indexes（D2：2.91；D3：2.48；D4：2.58）, the yellow indexes（D2：2.01；D3：1.21；D4：2.41）, the black indexes（D2：0.71；D3：0.53；D4：0.52）obviously changed, which showed no statistical difference, p>0.05.Conclusion:The established digital complexion inspection system can record the basic information of patients, symptoms and signs, laboratory results and facial image information; the system can effectively preprocess the input facial image and make the following operations, such as image translation, rotation, cutting, denoising, edge extraction, which can make the following consistent color research with collected facial images for different patients. The system can make RGB color analysis, HIS color analysis and the the five colors analysis with facial image of patients after pretreatment. The established digital complexion inspection system can make up the deficiency of the TCM diagnosis methods, which contributed to get rid of the interruptions of subjective factor on TCM diagnosis. And it also can provide objective and quantitative data for clinical facial inspection, which have some application value.Through the application of the established digital complexion inspection system the objective and quantitative study on the color of faces in CHB patients was carried on, and its result can reflect the evolution law of the energy attenuation of liver and spleen, which has certain reference value for clinical diagnosis and treatment and also contribute to improve the service ability and level of TCM.更多还原