The stress Electrocardiogram (ECG) gives more efficient results for the diagnosis of cardiovascular diseases, which may not be apparent when the patients are at rest. However, the noise produced by the movement of the patient and the environment often contaminates the ECG signal. Motion artifact is the most prevalent and difficult type of interference to filter in stress test ECG. It corrupts the quality of the desired signal thus reducing the reliability of the stress test. In this work, we first describe a quantitative study of adaptive filtering for processing the stress ECG signals. The proposed method uses the motion information obtained from a 3-axis accelerometer as a noise reference signal for the adaptive filter and the optimal weight of the adaptive filter is adjusted by the Modified Error Data Normalized Step-Size (MEDNSS) algorithm. Finally, the performance of the proposed algorithm is tested on the stress ECG signal from the subject.
