ALGORITHMS OF ECG FILTERING WITH DYNAMICALLY VARIABLE WINDOW SIZE

Abstract

Approximating and nonlinear locally adaptive filters with dynamically variable size of scanning window that have simple implementation for processing ECGs registered with sampling frequency 400 Hz are considered. Statistical estimates of filter quality according to MSE and SNR criteria are obtained. Adaptation parameters that depend on signal and noise level are analyzed as well as the corresponding plots of window variation for processing input signal. Adaptive filter output signals are presented for ECG with different noise levels. Recommendations on their application are given. Advantages of applied adaptation parameters are marked.

Keywords

• динамічні апроксимуючі алгоритми, нелінійні локально-адаптивні фільтри, електрокардіограма, електроміографічний шум, оцінки ефективності.