QRS detection based on hidden Markov modeling

An experiment is described to investigate the application of hidden Markov modeling (HMM) to QRS detection. Previously reported detection algorithms employ decision rule stages which require specification and adjustment of numerous parameters for adequate detection accuracy. The hidden Markov modeling approach provides automatic estimation of all parameters in the decision rule stage from each ECG file undergoing analysis. Preliminary results using a simple HMM detector show performance comparable to that of a multiparameter threshold based decision rule scheme.<
>