倒双谱估计在海洛因吸毒者脉象信号检测中的应用  

Application of Bicepstrum Estimation in Detecting Pulse Signals of Druggers

作　　者：黄镭[1] 刘宗行[1] 蔡坤宝[1] 

机构地区：[1]重庆大学通信工程学院,重庆400044

出　　处：《重庆工学院学报》2007年第13期98-102,共5页Journal of Chongqing Institute of Technology

基　　金：重庆市自然科学基金资助项目(CST;2004BB5061)

摘　　要：提出了用倒双谱估计和BP神经网络来检测中医脉象信号的方法,用该方法对15例吸毒者和15例正常人的脉搏波的倒双谱进行了估计,提取出每例脉搏波信号的倒双谱的3个特征参数,并将这3个参数构造1个三维特征参数向量作为所设计的BP神经网络的1个样本输入.研究结果表明,该方法能有效地区分吸毒者和正常人脉象信号的差异,并可进行自动识别.The bicepstnnn estimation and BP neural network are used to detect human pulse signals suecessfully. As the first step, the bicepstra of the pulse sills for 15 heroin addicts and 15 healthy persons are estimated. Then, three parameters are extracted from every bicepstrum which are integrated into a three-dimensional vector used as a sample input to a designed two-layer BP neural network. The research result shows that the method presented in this paper can effectively recognize the difference between the pulse signals of heroin druggers and normal persons, and thus can automatically identify.

关 键 词：脉象信号 高阶倒谱 倒双谱 BP神经网络 

分 类 号：R318.04[医药卫生—生物医学工程]