基于CEEMDAN的脑组织氧信号处理算法及其应用  

Brain Tissue Oxygen Signal Processing Algorithm and Application Based on CEEMDAN

作　　者：李泽熙 尹琦 康富豪 叶继伦 张旭 于辉 LI Zexi;YIN Qi;KANG Fuhao;YE Jilun;ZHANG Xu;YU Hui(Innovation Institute,Guangdong Biolight Meditech Co.,Ltd.,Zhuhai Guangdong 519080,China;School of BiomedicalEngineering,Shenzhen University Health Science Center,Shenzhen Guangdong 518000,China;Guangdong Key Laboratory ofBiomedical Signal Detection and Ultrasound Imaging,Shenzhen Guangdong 518000,China)

机构地区：[1]广东宝莱特公司创新研究院,广东珠海519080 [2]深圳大学医学部生物医学工程学院,广东深圳518000 [3]广东省生物医学信号检测与超声成像重点实验室,广东深圳518000

出　　处：《中国医疗设备》2024年第7期14-19,35,共7页China Medical Devices

基　　金：珠海市人才项目(2120004000207);深圳市科创委项目(JSGG20210713091811038)。

摘　　要：目的提出一种基于三波长的脑组织氧监测数学模型。方法结合完全集合经验模态分解与自适应噪声(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,CEEMDAN)和排列熵(Permutation Entropy,PE),对脑氧信号进行处理,同时用信噪比矫正PE值区间的选择,以自适应滤波提高脑血氧降噪。结果本文采集了3例志愿者数据并将其和现有市场设备进行对比,结果表明,与经验模态分解算法、集合经验模态分解算法相比,本文提出的CEEMDAN算法与对比设备的均方根误差均小于1.7,3组数据的皮尔逊相关系数分别为0.885、0.899、0.883,整体相关性较高(P<0.01);在降氧实验中,该算法可有效监测脑氧值变化趋势,具有较好的实用价值。结论该算法可有效去除基线漂移、低频噪声以及高频噪声,解决模态混叠和残余噪声问题,并提高滤波的准确性,进一步提高重构信号的信噪比,提升系统的有效性和稳定性。Objective To propose a mathematical model for brain tissue oxygen monitoring based on three wavelengths.Methods Combined complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)and permutation entropy(PE),the brain oxygen signal was processed.Meanwhile,the selection of PE value intervals was corrected by signal-tonoise ratio to improve cerebral oxygen noise reduction through adaptive filtering.Results In this paper,the data of three volunteers were collected and compared with the existing market equipment.The results indicated that compared to the empirical mode decomposition algorithm and ensemble empirical mode decomposition algorithm,the CEEMDAN algorithm proposed in this study had a root mean square error of less than 1.7 when compared to the reference equipment.The Pearson correlation coefficients for the three sets of data were 0.885,0.899,and 0.883 respectively,showing a high overall correlation(P<0.01).In hypoxia experiments,this algorithm could effectively monitor the trend of changes in brain oxygen values and demonstrated good practical value.Conclusion The algorithm effectively can remove baseline drift,low-frequency noise,and high-frequency noise,addressing mode mixing and residual noise issues and can enhance the accuracy of filtering,further improving the signal-to-noise ratio of the reconstructed signal and enhancing the effectiveness and stability of the system.

关 键 词：脑血氧 经验模态分解 排列熵 近红外光谱 三波长 

分 类 号：R197.39[医药卫生—卫生事业管理]