| Comparisons of six filters to remove high-frequency noises from interfered ECG |
|
Jong Cook Park, Sang Hyun Park, Hye Jin Kim |
|
Department of Anesthesia and Pain Medicine, Jeju National University School of Medicine, Jeju, Korea |
Correspondence:
Jong Cook Park, Email: pjcook@jejunu.ac.kr
|
|
|
| Abstract |
Background
An Electrocardiography obtained during operation is frequently contaminated with different types of noise. Noise sources such as an electrosurgical unit are obstacles that must be overcome by expert monitoring systems. We examined the six different low-pass filters (LPFs) designed to remove high frequency noises from an interfered ECG.
Methods
ECG signals were collected using anesthesia monitors at a sampling rate of 300 Hz from 15 patients under general anesthesia. A high frequency sinusoidal noise signal of 38 Hz was added to the ECG signals to simulate - interfered ECG signals, Six LPFs (Window-sinc, 0% 6-pole, 0.6% 6-pole, 6% 6-pole, 16% 6-pole, and 26% 6-pole Chebyshev filler) were designed to remove the high frequency noise from the interfered ECG signal were constructed and their performance were tested. The one-way ANOVA was used to do statistical analysis on the differences between the LPFs.
Results
In the correlation index and the reduction of magnitude of the high frequency noise, the Window-sinc filter was significantly better than the 6-pole Chebyshev groups (P< 0.01). In the 6-pole Chebyshev filters, rise in the ripple percentage increased noise reduction ratio but caused the distortion of QRS complex.
Conclusions
We suggest the Window-sinc filter which performs better than 6-pole Chebyshev filters for removing noise from ECG signals. Further works are required to implement an optimized filter in real-time on an expert monitoring system. |
| Key Words:
Electrosurgery; Electrocardiography; Low frequency filter; Noise; Signal processing |
|
PDF Links
Full text via DOI
Download Citation
