非接触式动态光电3D脉象信息检测及多维分析识别技术研究

项目名称： 非接触式动态光电3D脉象信息检测及多维分析识别技术研究

项目编号： No.81202856

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 医学八处

项目作者： 刘志臻

作者单位： 福建中医药大学

项目金额： 23万元

中文摘要： 脉诊是中医临床诊断极具特色及重要的依据之一，目前市场上使用的脉象仪多以接触式的压电传感器为主，可表现出脉象的幅度和频率，多属二维成像。但脉象形成极为复杂，应采用更先进的三维空间与四维时间的动态采集，并引入人工智能知识工程分析识别技术，对中医常见脉象（平脉、弦脉、滑脉、弦滑脉、细脉、细弦脉、细滑脉）的三维脉图进行"位数形势"多属性特征综合分析，建立对应信息检测模型。本研究是以3D激光非接触传感检测手部寸关尺脉博处的脉象信息，具有高精密的光学分辨率、多维性及灵敏性等优势。同时构建临床脉象云端信息传输储存知识库，以ICC值、Kappa值分析同一评分者评分的稳定性、不同评分者评分的一致性及重测信度，完成脉象信息特征赋值的标准化；采用神经网络与模糊逻辑方法对特征赋值实时分析识别归类。本研究不仅定位精度高还能揭示脉象更丰富的生理病理内涵，将为中医脉象的采集及信息特征分析提供新的研究思路和方法。

中文关键词： 脉象；光电；人工智能；云计算；物联网

英文摘要： Pulse analysis of human body is one of the major methods for the clinical diagnosis of TCM. Currently the instruments used in the field are mostly kinds of the contact approach, such as the piezoelectric sensors, usually these devices can only present the amplitude and frequency of the pulse, and only two dimensional curves could be shown. Pulse information is very complex, the study presents in a more advanced way, the presentation of 3D/4D pictures are shown, and the artificial intelligence technologies for the detailed analysis is utilized. TCM pulse includes flat pulse, string pulse, slide pulse, string flat pulse, thin pulse, thin string pulse, and thin slide pulse, etc. Based on the 3D non-contact laser sensor, the digitized pulse information and the analysis models could be established. The information of the three major pulses on the wrist could be detected in a high precision and high resolution through the electro-optical device. The knowledge base could be transformed and exchanged through the cloud technology. The ICC value and the Kappa value analysis are used for the analysis of the stability, consistency and reliability of the data which captured from different patients. The neural network and the fuzzy logic methods are used for the analysis for the correlation of the cause and fact, and provide

英文关键词： pulse condition；photoelectricity；artificial intelligence；cloud computing；internet of things