光谱相减方法用于无创生化检测中背景扣除实验研究  

Experimental Study of Differential Spectrum Method for Elimination of Tissue Background in Noninvasive Biochemical Detection

作　　者：丁海泉[1] 卢启鹏[1] 高洪智[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,应用光学国家重点实验室,吉林长春130033

出　　处：《光谱学与光谱分析》2012年第9期2347-2350,共4页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(60938002);应用光学国家重点实验室开放基金项目(Y1Q03FQ113);中国科学院知识创新工程领域前沿资助项目(Y00232Q100)资助

摘　　要：在近红外光谱无创生化检测中,血流容积光谱相减方法在理论上可以消除组织背景等干扰因素,提取出血液的有效光谱信息。为论证血流容积光谱相减方法有效性,设计了相应的模拟实验。采用生物分子水溶液模拟血清样品,不同吸收特性的滤光片模拟组织背景干扰,可变厚度样品池模拟血流容积变化。比较了光谱相减方法处理前后的模型精度,处理前模型定标相关系数(Rc)为0.476,交叉检验标准差(RMSECV)为437mg·dL-1;处理后Rc达到0.977,RMSECV降至301mg·dL-1。实验结果表明血流容积光谱相减方法能够较好抑制或扣除组织背景干扰,大幅度提高模型预测精度。In noninvasive biochemical detection,the differential spectrum method based on the change in blood volume can eliminate the interference of human tissue background in theory,and obtain effective spectrum information of blood.In order to demonstrate the effectiveness of the differential spectrum method,simulated experiment was designed.Biological molecules solutions were used for simulating serum sample,filters with different absorption characteristic were used for simulating interference of tissue background,and an adjustable path-length cell was used for simulating blood volume change.Model accuracies of pre-and post-treatment with differential spectrum method were compared.Thus treated,the root mean square error of cross validation（RMSECV） reduced from 437 to 301 mg·dL-1.The experimental results indicate that using the differential spectrum method can effectively restrain the interference of tissue background,and greatly improve the prediction precision of calibration model.

关 键 词：近红外光谱 光谱相减方法 无创生化检测 模拟血清 

分 类 号：O433.4[机械工程—光学工程]