文章摘要

文章信息/Info

Title:: Design and implementation of biometric signal purification algorithm

Author(s):: SUN Xiang1; HUA Gang2; 1. Information Department of Affiliated Hospital of Jiangsu University,Zhenjiang 212000,Jiangsu，China;2. School of Information and Control Engineering,China University of Mining and Technology,Xuzhou 221116,Jiangsu，China

Keywords:: Biology; Characteristic signal; Signal purification; Matched filtering; Detector; Time-frequency distribution

摘要:: 生物特征信号在频域内很容易受到干扰，固定阙值的提纯算法在实际应用中，受到多类干扰信号影响，导致提纯困难。本设计基于自适应时频分解的生物特征信号提纯算法，构建生物信号采集模型，对生物信号在时域和频域内进行特征分解，得到单分量特征。设计匹配滤波检测器估计出生物信号频域特征系数，将解析信号代入滤波器中，进行实时滤波处理，得到生物特征信号的基本分量。用提取出的信号分量的时频分布进行时域信号的重建，获得提纯后的生物信号分量。实验结果表明，在强噪声干扰下，所设计的提纯算法具有较好的去噪和检测性能。与传统提纯算法相比，检测效率提高了8.5%以上，为生物复杂信号的研究奠定了理论基础。

Abstract:: Biometric signals are easily disturbed in the frequency domain, and the purification algorithm with fixed threshold is affected by many kinds of interference signals in practical applications, which leads to the difficulty of purification.A biometric signal purification algorithm based on adaptive time-frequency decomposition was designed, and a biometric signal acquisition model was constructed. The biometric signal was decomposed in time domain and frequency domain to obtain single component features.A matched filter detector was designed to estimate the frequency domain characteristic coefficients of bio-signal. The analytic signal was substituted into the filter for real-time filtering and the basic components of bio-signal were obtained. The time-frequency distribution of the extracted signal components was used to reconstruct the time-domain signal, and the purified biological signal components were obtained. Experimental results show that the proposed algorithm has better performance in denoising and detection under strong noise interference. Compared with the traditional purification algorithm, the detection efficiency is increased by more than 8.5%, which lays a theoretical foundation for the study of complex biological signals.