Iliev, Ivo Tsvetanov; Krasteva, Vessela Tzvetanova; Tabakov, Serafim Dimitrov

Real-time detection of accidental pathologic cardiac events in the electrocardiogram

2006

Proceedings of the Technical University – Sofia, 56(1), 2006, pp. 206-211

TU – Sofia

1311-0829 [issn]

Article

eng

The major efforts for improving the automated cardiac diagnostics are directed towards developments of innovative hardware and software solutions for computer-assisted electrocardiogram (ECG) monitoring systems. The presented work describes a method for fast detection of pathologic cardiac events by real-time ECG analysis. It is convenient for embedding in microcontroller-based autonomous system for monitoring of high-risk cardiac patients. The algorithm involved on-line operating procedures, including preprocessing filtration, threshold-based QRS detection, interbeat RR-intervals analysis and QRS pattern waveform analysis. Aiming at a simple solution, we adopted specific strategies for signal processing acceleration and for reduction of the operational memory size, such as the resolution reduction of the QRS pattern waveform. Moreover we implemented simplified techniques for rating of the similarity between the QRS pattern of the tested beat and the accumulated QRS pattern of the preceding heartbeats. The repetition of similar QRS pattern waveforms combined with small variances of the RR-intervals, was interpreted as a normal rhythm. However, the appearance of a number of deviations either from the mean RR interval, or from the cumulative QRS pattern waveform, was detected as a sustained pathologic event. The developed algorithm was implemented in Matlab environment. It was tested with internationally recognized ECG databases. Several examples are presented and discussed.

Reviewer prof. Ivan Dotsinsky. – Figuras, formulas, schemes. - References, pp. 211