ECG Signals Denoising and Features Extraction by Applying UFIR Smoothing with Optimal $q$-Lag in the State Space

The medical sector requires efficient applications for the extraction of features in ECG signals. This is required to detect diseases and abnormalities presented in the heart. Some works consider filters based on the wavelet transform as a standard technique. Here, we develop an unbiased finite impulse response filter in the state space for ECG signals with optimal q - lag which outperforms to wavelet-based filter techniques. Where several simulations are compared using different types of mother wavelets with Gaussian white noise with the UFIR filter. The experiments are based on the MIT-BIH database using records in normal and atrial fibrillation condition. Also, UFIR smoothing is applied to remove motion in ECG signals caused by artifacts. Finally, temporal features extraction of the ECG signals expressed by normalized histograms is performed. According to the criteria of the gold standard, the results provided by the UFIR filter demonstrate a significant separation between studied pathologies.