Motion artifact removal in ECG signals using multi-resolution thresholding

The electrocardiogram (ECG) is a powerful non-invasive tool which allows for diagnosis of a wide range of heart conditions. Today, portable ECG recording devices, equipped with a transmitter, can be used to provide health related information and to trigger alarms in case of life threatening situations. However, these devices suffer from motion induced artifacts. While much research has been conducted to remove time invariant noise, the removal of motion induced artifacts remains an unsolved problem. We therefore introduce a new method which removes these artifacts. This is done by obtaining an estimate of the artifacts using the stationary wavelet transformation. An automatic multi-resolution thresholding scheme which uses a robustified QRS detection is proposed. Real data examples as well as simulations are given which illustrate the performance of the method.