WAVELET METHODS FOR CARDIO SIGNALS PROCESSING

Abstract

The article discusses the use of wavelet transforms and algorithms in the digital processing of cardiosignals in a number of medical applications due to their good adaptability to the analysis of non-stationary signals (that is, those whose statistical characteristics change with time). Since an electrocardiogram is a transient signal, wavelet methods can be used to recognize and detect key diagnostic features. In real-time systems for digital signal processing, it is important that mathematical operations are performed quickly, and the time required to execute commands must be known precisely and in advance. For this, both the program and the hardware must be very effective. In digital signal processors, the most important mathematical operation and the core of all digital signal processing algorithms is multiplication, followed by summation. Fast execution of the multiplication operation followed by summation is very important for implementing real-time digital filters, signal processing, matrix multiplication, and graphic image manipulation. Therefore, all this requires the need to improve methods, algorithms and signal processing programs that determine the quality and performance of digital systems. The technique is based on the transformation of the heart rate to simple harmonic oscillations (fast Fourier transform, autoregressive analysis) with different frequencies. In this case, the sequence of heartbeats is converted into a power spectrum of fluctuations of the duration of the RR intervals, which are a sequence of frequencies characterizing HRV. Most often, the area bounded by the spectral power curve corresponding to a certain defined frequency range, that is, the power within a limited frequency range, is estimated.