The Application Of Higher-Order Cepstra To The Analysis Of Human Pulse Signals

The unique diagnostic methods and particularly curative effects of traditional Chinese medicine are obvious to all. Along with the development of sensors and computer technology, people bend themselves to the objectivity of the Chinese medicine, and hope to apply the modern technologies and instruments to push the modernization of the traditional Chinese medicine, which is the main motive in this paper.Higher-Order Spectra is an advancing research aspect in signal processing field, which is the main tool for analyzing non-Gaussian signals. Bicepstrum is the most common method of Higher-Order Spectra mathematical analysis, In this paper, the algorithm of bicepstrum approach is used to analyze the bicepstrum estimation of pulse signals for 15 heroin addicts and 15 healthy persons to classify the pulse signals. As the magnitude of bicepstrum estimation is only exist at first-quadrant,third- quadrant and diagonal , so we use the diagonal slice of bicepstrum to analyze. it is found that the zero-magnitude, covariance in a specific region and the radio between the signal energy of a specific region of bicepstrum's diagonal slice and the total original signal energy of heroin addicts is generally higher than that of healthy persons. Using this three parameters as characteristic parameters, three critical parameters is determined that is used to classify heroin addicts and healthy persons. Thus, as to zero-magnitude, 14 healthy persons and 14 heroin addicts are identified, Only one heroin addict B13 and one healthy persons Z15 is misjudged, to covariance in a specific region, two heroin addict B13,B14 and one healthy persons Z15 is misjudged,and to the radio between the signal energy of a specific region and the total signal energy, one heroin addict B06 and one healthy persons Z15 is misjudged. As the computation number of bicepstrum is very large and the advantage of 1 12 dimension cepstrum, so 1 12 dimension cepstrum is used to classify the pulse signals. We extract the zero-magnitude, information entropy in a specific region and the radio between the energy of cepstrum in a particular range and that in another range, as to zero-magnitude , 13 healthy persons and 14 heroin addicts are identified, Only one heroin addict B13 and two healthy persons Z01 and Z10 is misjudged, to information entropy, 14 healthy persons and 13 heroin addicts are identified,Only one healthy Z13 and two heroin addicts B03 and B06 is misjudged,to the radio between the energy of cepstrum in a particular range and that in another range,all of the 15 healthy persons are identified,but heroin addicts B11 B12 and B13 are misjudged.Finally,we organize the three characteristic parameters from above method as characteristic vector,adopts the improved BP Algorithm, i.e. LMAlgorithm.The experiments show the apply of LMBP on human pulse has quick trained rate and high discrimination which reachs 100% and 96.7% respectively. It is shown that at the aspect of extracting characteristic information of pulse signals, the bicepstrum estimation and 1 12 dimension cepstrum estimation all have the higher discrimination. But the computation number of 1 12 dimension cepstrum is fewer than that of bicepstrum while the discrimination of 1 12dimension cepstrum estimation is also descend.This paper represents the basic conceptions and theories of bicepstrum and 1 12dimension cepstrum estimation in detail.While using the algorithm of bicepstrum and 1 12 dimension cepstrum to analyze pulse signals, this paper also derives, verifies and uses them.