Wavelet-based pulse-abnormality analysis for heroin addicts / 生物医学工程学杂志

ABSTRACT

Using wavelet transforms method, time-frequency characteristics of pulse signals from 15 heroin addicts and 15 healthy persons were analyzed. According to 3-D and contour plots used to display discrete dyadic wavelet transforms, the significant difference of time-frequency characteristics between the signals of heroin addicts and healthy persons were revealed. A primary criterion was also obtained,with the criterion, 15 heroin addicts were entirely identified, while two healthy subjects were misidentified. The research result shows that the wavelet-based multiresolution analysis is a very effective method to extract characteristics of pulse signals. It is valuable to the diagnosis and therapy for heroin addicts.