Journal of Physics Conference Series · 2011 · 12 citations · 6 references
MedicineBiomedical EngineeringCardiologyBiomedical Signal ProcessingClinical CardiologyElectrophysiologyCardiac ArrhythmiasCardiac PhysiologyCardiac Signal ProcessingElectrocardiogram AnalysisElectrocardiographyVital SignsArrhythmia DetectionBeat ClassificationCardiac Feature ExtractionEcg Beat ClassificationEcg Feature ExtractionElectrocardiogram Feature ExtractionHolter Electrocardiogram AnalysisQrs Complex Detection
In this work a detection and classification algorithm for heartbeats analysis in Holter records was developed. First, a QRS complexes detector was implemented and their temporal and morphological characteristics were extracted. A vector was built with these features; this vector is the input of the classification module, based on discriminant analysis. The beats were classified in three groups: Premature Ventricular Contraction beat (PVC), Atrial Premature Contraction beat (APC) and Normal Beat (NB). These beat categories represent the most important groups of commercial Holter systems. The developed algorithms were evaluated in 76 ECG records of two validated open-access databases "arrhythmias MIT BIH database" and "MIT BIH supraventricular arrhythmias database". A total of 166343 beats were detected and analyzed, where the QRS detection algorithm provides a sensitivity of 99.69 % and a positive predictive value of 99.84 %. The classification stage gives sensitivities of 97.17% for NB, 97.67% for PCV and 92.78% for APC.
6
A Real-Time QRS Detection Algorithm
Jiapu Pan, W.J. Tompkins · IEEE Transactions on Biomedical Engineering · 1985 · 7.7K citations