Conference Paper

Combining algorithms in automatic detection of R-peaks in ECG signals

Philips Res. Lab., Aachen, Germany
DOI: 10.1109/CBMS.2005.43 Conference: Computer-Based Medical Systems, 2005. Proceedings. 18th IEEE Symposium on
Source: IEEE Xplore


R-peak detection is the crucial first step in every automatic ECG analysis. Much work has been carried out in this field, using various methods ranging from filtering and threshold methods, through wavelet methods, to neural networks, and others. Performance is generally good, but each method has situations where it fails. In this paper we suggest an approach to automatically combine different algorithms, here the Pan Tompkins and wavelet algorithms, for detection of R-peaks in ECG signals, in order to benefit from the strengths of both algorithms. Experimental results and analysis are provided on the MIT-BIH Arrhythmia Database. We obtained substantial improvements on the test data with respect to the best individual algorithm.

4 Reads
  • Source
    • "Wavelet analysis, continuous or discrete, has been applied to ECG signals, among many other purposes [21], to obtain the heart rate. The more recently developed wavelet based algorithms [22] [23] overcome some of the drawbacks of the classical detection algorithms [24] "
    [Show abstract] [Hide abstract]
    ABSTRACT: This work presents a novel easy-to-use system intended for the fast and non-invasive monitoring of the Lead I electrocardiogram (ECG) signal by using a wireless steering wheel. The system uses a dual ground electrode configuration connected to a low-power analog front-end to reduce 50/60 Hz interference and it is able to show a stable ECG signal with good enough quality for monitoring purposes in less than 5 s. A novel heart rate detection algorithm based on the continuous wavelet transform (CWT) has been implemented, which is specially designed to be robust against the most common sources of noise and interference present when acquiring the ECG in the hands, i.e., electromyographic (EMG) noise and baseline wandering. The algorithm shows acceptable performance even under non-ordinary high levels of EMG noise and yields a positive predictivity value of 100.00 % and a sensitivity of 99.75 % when tested in normal use with subjects of different age, gender and physical condition.
    IEEE Sensors Journal 03/2012; 12(99-PP):1 - 1. DOI:10.1109/JSEN.2011.2118201 · 1.76 Impact Factor
  • Source
    • "Martínez et al also use wavelet transforms, but don't claim real-time results. Fernández et al [6] present some experimental findings from methods which combine filtering, thresholding and wavelet algorithms. Benitez et al [1] used Hilbert transforms, which they report has good results in the presence of noise. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes a method for the detection of QRS complexes. The method applies the algorithm proposed by Kupeev [K.Y. Kupeev, On significant maxima detection: a fine-to-coarse algorithm, in: Proceedings of the 13th International Conference on Pattern Recognition, IEEE, vol. 2, August 1996, pp. 270–274] to a shifting window over the ECG signal. The method was tested on the MIT-BIH Arrhythmia database and achieves a total failure rate of 1.7%, and processes a single record (30 min sampled at 360 Hz) within 1.5 s. The method requires little or no pre-processing. Further, the method makes very few a priori assumptions about the nature of the signal.
    Biomedical Signal Processing and Control 04/2006; 1(2):169-176. DOI:10.1016/j.bspc.2006.08.002 · 1.42 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Widespread use of medical information systems and explosive growth of medical databases re-quire methods for efficient computer assisted analysis. In the paper we focus on the QRS complex detection in electrocardiogram but, the idea of further recognition of anomalies in QRS complexes based on the im-munology approach is described, as well. In order to detect QRS complexes the neural network ensemble is proposed. It consists of three neural networks. The details referring to this solution are described. The results of the experimental study are also shown.
Show more