A piecewise geometric analysis method for real-time ambulatory ECG detection
Abstract
BACKGROUND: As one of the pervasive healthcare services, Ubiquitous cardiac care (UCC) systems should have at least two significant characteristics: real-time detection capability for cardiac arrhythmia events and a small resource requirement for its computation and storage.
PURPOSE: Due to the strict-constrained system support and ambulatory signal quality in the out-of-hospital pervasive healthcare applications, a dedicated real-time AED (Ambulatory Electrocardiograph Detection) algorithm has been implemented.
METHODOLOGY: By adopting the piecewise geometric analysis method, this algorithm can provide a real-time continuous detection capability for QRS complexes, which consists of three main functional modules: the Data preparation; the R-wave vertex discovery; and the QRS complex recognition. Currently, this algorithm has been applied on an on-line UCC application system at the hospital for more than 30 patients.
RESULT: The performance evaluation has been made not only on the standard MIT-BIH cardiac arrhythmia database but also on the clinical testing. The experiential results explore this algorithm has in average sensitivity of 99.37% and specificity of 99.72%.
CONCLUSION: This AED algorithm has minimal beat detection latency and a less computation consumption, which make it meet the requirements of ubiquitous cardiac care applications.