Filtering of noise in electrocardiographic signals using an unbiased and normalized adaptive artifact cancellation system

The electrocardiogram (ECG) is routinely used for the diagnosis of cardiovascular diseases. The removal of artifacts in ambulatory ECG recordings is essential in many biomedical applications. In this paper, we present the design of an unbiased linear filter with normalized weight coefficients in an adaptive artifact cancellation (UNAAC) system. We also develop a new weight coefficient adaptation algorithm that normalizes the filter coefficients, and utilize the steepest-descent algorithm to effectively cancel the artifacts present in ECG signals. The proposed UNAAC system was tested through experiments on the benchmark MIT-BIH database. Empirical results demonstrate that the UNAAC system can effectively eliminate two types of predominant artifacts: baseline wander and muscle-contraction artifact. Furthermore, the proposed UNAAC system achieved significantly higher signal-to-noise and signal-to-error ratios in the enhanced ECG signals, as compared with the normalized least-mean-square (NLMS) filter.