A realistic coupled nonlinear artificial ECG, BP and respiratory signal generator for assessing noise performance of biomedical signal processing algorithms

Journal article

Extensions to a previously published nonlinear model for generating realistic artificial electrocardiograms to include blood pressure and respiratory signals are presented. The model accurately reproduces many of the important clinical qualities of these signals such as QT dispersion, realistic beat to beat variability in timing and morphology and pulse transit time. The advantage of this artificial model is that the signal is completely known (and therefore its clinical descriptors can be specified exactly) and contains no noise. Artifact and noise can therefore be added in a quantifiable and controlled manner in order to test relevant biomedical signal processing algorithms. Application examples using Independent Component Analysis to remove artifacts are presented.

Authors: Clifford, G; McSharry, P

• Journal: Proceedings of SPIE - The International Society for Optical Engineering
• Volume: 5467
• Pages: 290-301
• Publication date: 2004-01-01
• DOI: 10.1117/12.544525
• ISSN: 0277-786X
• Language: English
• Keywords: Respiration; EKG; Blood pressure; Independent component analysis; Noise; Measurement error; Noise removal; Artificial signal; Electrocardiogram; Nonlinear model