PPG多频域特征的改进灰狼身份识别方法  

Improved grey wolf identification method based on PPG multiple frequency⁃domain features

作　　者：朱志敏 陈小惠 陈勤达 宋玲玉 ZHU Zhimin;CHEN Xiaohui;CHEN Qinda;SONG Lingyu(College of Automation and Artificial Intelligence,Nanjing University of Posts and Telecommunications,Nanjing 210046,China)

机构地区：[1]南京邮电大学自动化学院、人工智能学院,江苏南京210046

出　　处：《现代电子技术》2023年第15期71-75,共5页Modern Electronics Technique

摘　　要：针对目前生物信息身份识别方法大多存在识别率不高,或出现特征提取方式数量不够、特征间关联性较低等问题,提出一种PPG多频域特征的改进灰狼身份识别方法。针对现有方法特征提取数量不足,对光电容积脉搏波(PPG)信号进行多频域多特征提取,即从时域、频域、小波域三个域度共提取20维特征;为保留特征间的相关性,利用改进的FCBF算法对提取到的20维特征进行降维,获得特征子集;为保证最终识别的准确率,构建改进灰狼算法优化支持向量机(IGWO⁃SVM)的分类模型,对降维后的特征集进行训练和测试,最终完成PPG身份识别。仿真实验结果表明,该方法的识别准确率可达到98.61%。Currently,most of the biometric identification methods have low recognition rates,insufficient feature extraction means and low feature correlation.In view of this,an improved gray wolf identification method based on PPG(Photoplethysmography)multiple frequency⁃domain features is proposed.In order to increase the quantity of extracted features,the PPG signal is extracted based on multi⁃frequency domain and multi⁃features,that is,20⁃dimensional features are extracted from PPG signals in time domain,frequency domain and wavelet domain.In order to retain the correlation among features,the improved FCBF(fast correlation⁃based filter)algorithm is used for dimensionality reduction of the extracted 20⁃dimensional features,so as to obtain a feature subset.To ensure the accuracy of the final identification,an IGWO⁃SVM classification model(SVM optimized by improved gray wolf algorithm)is constructed to train and test the reduced feature set for PPG identification.Simulation results show that the proposed method can achieve an identification accuracy of 98.61%.

关 键 词：PPG信号 FCBF算法 IGWO⁃SVM算法 识别率 降维 身份识别 

分 类 号：TN911.73-34[电子电信—通信与信息系统] R318[电子电信—信息与通信工程]