Cancellation of high-frequency noise in ECG signals using adaptive filter without external reference

Abstract

ECG recordings are often contaminated by high-frequency noises, such as power-line interference, electromyography (EMG) noise, and instrumentation noise. The use of adaptive filters in cancelling the noise requires an external reference to estimate the noise and, in turn, subtracting it from the noisy ECG. However, this is often ineffective due to the fact that the reference signal cannot be well-correlated with the noise part in the primary input. An adaptive structure used to deal with the noise without the need of external reference is addressed in this paper. Basic idea behind this is due to the fact that ECG signals can be treated as periodic signals in some frequencies. This makes it possible for an adaptive filter to estimate the clean ECG signal from the noisy one. The usefulness of this approach is confirmed by using simulated ECG as well as real ECG from the database of MIT-BIH.