基于插值的准周期电生理信号样本构建的研究

Simultaneously periodic sample's establishment for Quasi-periodic signal based on the interpolation

人体是一个复杂的有机体,同一类型的生理数据来自不同的个体,要对这些来自不同信号源的数据进行比较,需要建立统一的标准,获得优化的数据样本,并能消减不同信号源的差异以及随机差异,形成具有可比性的多组生理信号,才能进一步有效地进行研究。在复杂的生理信号中存在着一类具有的周期性的信号,它们之间存在着一定的共性,如脉搏和呼吸信号等,称为准周期电生理信号,它们的预处理过程有不少共同之处。因此,以这几种生理信号为主要对象,研究对该类生理信号数据进行融合前的数据预处理过程,研究并提出了基于插值理论构建准周期生理信号的等周期数据样本的方法。

Because of the complexity of human body as an organism, the same type of physiological data comes from different individuals, to compare this data from different signal sources, a unified standard must be established to acquire optimized data sample, narrow the differences of different signal sources and random difference, and form signals, which can effectively be applied to machine learning, and information fusion and so on. In the complex physiological signals, exist certain signals with periodic characteristics. They have certain, like cardiogram, pulse and respiration signals, also dubbed quasi-periodic electrophysiology signal, and can be pre-treated commonly. Therefore, based on these kinds of physiological signals, this paper studies the data preprocess of Quasi-periodic signal and gives the engineering calculation and implementation method.