基于稀疏重建的EMD脉搏色素密度谱的处理

发布时间：2018-07-12 15:04

本文选题：脉搏色素光谱 + 经验模态分解　；参考：《科学技术与工程》2015年05期

【摘要】：基于脉搏血氧计的吲哚青绿(Indocyanine green,ICG)脉搏密度谱测量方法是无创肝储备功能的主要检测原理。近红外脉搏色素密度谱的提取及处理直接影响着肝储备功能检测参数的计算精度。针对自行研发的脉搏色素密度谱采集装置采集的近红外光谱信号的光谱噪声及干扰问题,采用了结合稀疏二进制重建的经验模态分解法对ICG密度谱进行处理。将处理后提取的ICG密度谱进行分析计算,得到的ICG血浆消失率值与报告值相比较,其最大相对误差为5.1%;而平均相对误差为3.2%。实验结果表明,结合自适应稀疏二进制重构的经验模态分解方法对消除色素光谱的噪声是有效的,有效地提高了脉搏色素密度谱的测量精度,从而提高肝储备功能检测参数的精度。为近红外脉搏色素光谱提取优化提供了一种新方法,为抗干扰的无创肝储备功能的检测提供理论依据和实用方法。
[Abstract]:Indocyanine green (ICG) pulse density spectrum measurement based on pulse oximeter is the main detection principle of noninvasive liver reserve function. The extraction and processing of near infrared pulse pigment density spectrum directly affect the calculation accuracy of liver reserve function parameters. Aiming at the problem of spectral noise and interference of the near-infrared spectral signals collected by the pulse pigment density spectrum acquisition device developed by ourselves, the empirical mode decomposition method combined with sparse binary reconstruction is used to process the ICG density spectrum. The ICG density spectrum extracted after treatment was analyzed and calculated. Compared with the reported ICG plasma disappearance rate, the maximum relative error was 5.1 and the average relative error was 3.2. The experimental results show that the empirical mode decomposition (EMD) method combined with adaptive sparse binary reconstruction is effective in eliminating the noise of the pigment spectrum and improving the measurement accuracy of the pulse pigment density spectrum. In order to improve the accuracy of liver reserve function detection parameters. It provides a new method for the extraction and optimization of near infrared pulse pigment spectrum and provides a theoretical basis and a practical method for the detection of anti-interference noninvasive liver reserve function.
【作者单位】：吉林大学仪器科学与电气工程学院;长春理工大学生命科学技术学院;
【基金】：国家重大科学仪器设备开发专项的子题(120111222867002) 国家自然科学基金项目(21207047) 高等学校博士学科点专项科研基金课题(20120061110092) 吉林省科技发展计划重点项目(20120328)资助
【分类号】：R575

【参考文献】