【摘要】 小波变换具有良好的时频局域化性质 ,适合于非平稳信号的处理。脉诊是中医诊察疾病的重要手段 ,脉象反映的是人体的生理与病理信息 ,脉象信号具有随机性和非线性等特点。应用连续小波变换分析了 15例海洛因吸毒者和 15例正常人的脉象信号 ,提取了吸毒者脉象信号中的异常信息 ,为戒毒治疗的评估与改进提供客观依据。在尺度 0 .1、0 .2、0 .3和 0 .4下 ,对每一例脉象信号分别进行了处理 ,发现在 0 .2～ 0 .4s时间间隔内 ,海洛因吸毒者和正常人脉象信号的连续小波变换系数间存在显著差异 ,并提出了用于划分海洛因吸毒者与正常人的临界参数。研究结果表明 ,连续小波变换是处理脉象信号的有效方法更多还原

【Abstract】 It is well known that wavelet transforms have good time-frequency localization, which makes the transforms be especially suitable for analyzing nonstationary signals. Human pulse diagnosis is an important means for disease diagnosis in traditional Chinese medicine. Human pulse contains the physiological and pathological information of human bodies. Pulse signals characterize us randomness, nonlinearity and nonstationarity. Pulse signals of 15 heroin addicts and 15 normal persons are analyzed using a continuous wavelet transform. Thus the abnormal information containing in pulse signals of heroin addicts is extracted, which provides an objective means of evaluation and improvement for heroin withdrawal treatment. Processing every pulse signal under scales 0.1, 0.2, 0.3 and 0.4, it is fourd that the significant difference of wavelet transform coefficients in the time interval 0.2～0.4 seconds exists between heroin addicts and normal persons. A critical parameter used to classify heroin addicts and normal persons is given in this paper. The research result of this paper shows that the continuous wavelet transform is really an effective method for processing pulse signals.更多还原