人工神经网络-红外光谱法对人体血糖的无创监测分析  

Analysis of Noninvasive Measurement of Human Blood Glucose with ANN-NIR Spectroscopy

作　　者：左平[1,2] 李映红[3] 韩笑[1] 马驷良[1] 

机构地区：[1]吉林大学数学学院,长春130012 [2]空军航空大学基础部,长春130022 [3]长春大学理学院,长春130022

出　　处：《吉林大学学报（理学版）》2008年第6期1131-1136,共6页Journal of Jilin University:Science Edition

基　　金：国家科技“十五”公关项目基金(批准号:2004BA210A09);吉林大学“985工程”项目基金

摘　　要：针对人体血糖无创监测的近红外光谱分析，提出一种血液红外吸收光谱分析的新技术，通过对正常人的全血和血清与高糖全血的红外吸收光谱进行测量，应用人工神经网络BP算法，以16个特征波长处的吸收值作为网络特征参数，进行网络训练．结果表明，网络训练11次即可达到误差精度要求，误差在-0．01—0．03mmol／L，达到了国家计量检测规程关于生化分析的要求．Based on the near infrared absorption spectrum analysis of noninvasive blood glucose, this paper puts forward a new technique for the infrared absorption spectrum analysis of blood. By means of the measurement of infrared absorption spectra of the normal people＇ s whole blood, serum and high glucose＇ s whole blood, application of the artificial neural network BP arithmetic, and with the absorption values of sixteen characteristic wavelengths as the network characteristic parameters, the network training was carried out. Only repeating 11 times, can the network training reach the error precision required. The error between - 0.01 - 0.03 mmol/L fulfils the requirement of national metrological verification regulation on biochemical analysis. This research result shows valuable prospects for blood infrared absorption spectrum analysis and disease diagnoses.

关 键 词：红外吸收光谱 人工神经网络 无创血糖检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]