Feature extraction of pulse diagnosis signal based on Hilbert yellow transform

Abstract

Listen

Translate article

Based on the pulse theory of the traditional Chinese medicine, the pulse diagnostic instrument can diagnose diseases. The most important step in pulse diagnosis is to extract the characteristic points from the measured pulse wave. The flat pulse waveform is gained from the pulse wave measured by the photoelectric sensor. The collected signal is filtered and denoised through wavelet transform to make the waveform smoother and reduce the impact of noise on subsequent feature extraction. Feature extraction of signal is mainly analyzed and processed in time domain, frequency domain and time frequency domain. In the time domain, some feature points of the original waveform are extracted. In the frequency domain, Fourier transform of the original signal is used to extract the spectral characteristics of the signal. In the time-frequency domain, the feature of signals are extracted by using of the Hilbert yellow transform. The original signal is decomposed through EMD (Empirical Mode Decomposition) to obtain several IMF (Intrinsic Mode Function), and then the Hilbert transform is carried out on several IMF to obtain the Hilbert spectrum. All the spectra are summarized to obtain the original spectrum. Through the combination of the wavelet transform and Hilbert yellow transform to process the original waveform, the characteristics of the pulse wave can be more clearly and accurately in the time domain, frequency domain and time-frequency domain, which is conducive to the subsequent judgment of the disease corresponding to the waveform.