A new statistical PCA–ICA algorithm for location of R-peaks in ECG
The success of ICA to separate the independent components from the mixture depends on the properties of the electrocardiogram (ECG) recordings. This paper discusses some of the conditions of independent component analysis (ICA) that could affect the reliability of the separation and evaluation of is...
Saved in:
Published in | International journal of cardiology Vol. 129; no. 1; pp. 146 - 148 |
---|---|
Main Authors | , , |
Format | Journal Article |
Language | English |
Published |
Shannon
Elsevier Ireland Ltd
16.09.2008
Elsevier Science |
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The success of ICA to separate the independent components from the mixture depends on the properties of the electrocardiogram (ECG) recordings. This paper discusses some of the conditions of independent component analysis (ICA) that could affect the reliability of the separation and evaluation of issues related to the properties of the signals and number of sources. Principal component analysis (PCA) scatter plots are plotted to indicate the diagnostic features in the presence and absence of base-line wander in interpreting the ECG signals. In this analysis, a newly developed statistical algorithm by authors, based on the use of combined PCA–ICA for two correlated channels of 12-channel ECG data is proposed. ICA technique has been successfully implemented in identifying and removal of noise and artifacts from ECG signals. Cleaned ECG signals are obtained using statistical measures like kurtosis and variance of variance after ICA processing. This analysis also paper deals with the detection of QRS complexes in electrocardiograms using combined PCA–ICA algorithm. The efficacy of the combined PCA–ICA algorithm lies in the fact that the location of the R-peaks is bounded from above and below by the location of the cross-over points, hence none of the peaks are ignored or missed. |
---|---|
Bibliography: | SourceType-Other Sources-1 content type line 63 ObjectType-Correspondence-1 |
ISSN: | 0167-5273 1874-1754 |
DOI: | 10.1016/j.ijcard.2007.06.036 |