Threshold based system of noise detection and elimination for ECG signal

The Cardiovascular disease (CVDs) is one of the most common cause of death in the world. The analysis of Electrocardiograms (ECGs) is a important tools in early diagnosis arrhythmias. However sometime these measurement data would be corrupted by noises which may cause by the wrong equipment operation, poor contact of the electrode, or even the breath of the patients. These noises would make cardiologists or automatic CVDs detection system hard to make a correct diagnosis. Therefor, the noise detection and elimination from ECG data become an important project on Health Information System (HIS). In this study, we propose six most common types of noise. For each noise type detection, we apply difference signal preprocessing. If there exist some noise segments that have no information and can not be repaired, we will eliminate them and combine the remain usable segments into a complete signal.