============= ECG Detectors

A collection of 7 ECG heartbeat detection algorithms implemented in Python. Developed in conjunction with a new ECG database: http://researchdata.gla.ac.uk/716/. This repository also contains a testing class for the MITDB and the new University of Glasgow database. In addition the module hrv provides tools to analyse heartrate variability.

Installation

Linux / Mac::

pip3 install py-ecg-detectors [--user]

Windows::

pip install py-ecg-detectors [--user]

From source::

python3 setup.py install [--user]

Use the option --user if you don't have system-wise write permission.

ECG Detector Class Usage

Before the detectors can be used the class must first be initalised with the sampling rate of the ECG recording:

.. code-block:: python

from ecgdetectors import Detectors detectors = Detectors(fs)

See usage_example.py for an example of how to use the detectors.

Hamilton

Implementation of P.S. Hamilton, “Open Source ECG Analysis Software Documentation”, E.P.Limited, 2002. Usage::

r_peaks = detectors.hamilton_detector(unfiltered_ecg)

Christov

Implementation of Ivaylo I. Christov, “Real time electrocardiogram QRS detection using combined adaptive threshold”, BioMedical Engineering OnLine 2004, vol. 3:28, 2004. Usage::

r_peaks = detectors.christov_detector(unfiltered_ecg)

Engelse and Zeelenberg

Implementation of W. Engelse and C. Zeelenberg, “A single scan algorithm for QRS detection and feature extraction”, IEEE Comp. in Cardiology, vol. 6, pp. 37-42, 1979 with modifications A. Lourenco, H. Silva, P. Leite, R. Lourenco and A. Fred, “Real Time Electrocardiogram Segmentation for Finger Based ECG Biometrics”, BIOSIGNALS 2012, pp. 49-54, 2012. Usage::

r_peaks = detectors.engzee_detector(unfiltered_ecg)

Pan and Tompkins

Implementation of Jiapu Pan and Willis J. Tompkins. “A Real-Time QRS Detection Algorithm”. In: IEEE Transactions on Biomedical Engineering BME-32.3 (1985), pp. 230–236. Usage::

r_peaks = detectors.pan_tompkins_detector(unfiltered_ecg)

Stationary Wavelet Transform

Implementation based on Vignesh Kalidas and Lakshman Tamil. “Real-time QRS detector using Stationary Wavelet Transform for Automated ECG Analysis”. In: 2017 IEEE 17th International Conference on Bioinformatics and Bioengineering (BIBE). Uses the Pan and Tompkins thresolding method. Usage::

r_peaks = detectors.swt_detector(unfiltered_ecg)

Two Moving Average

Implementation of Elgendi, Mohamed & Jonkman, Mirjam & De Boer, Friso. (2010). "Frequency Bands Effects on QRS Detection" The 3rd International Conference on Bio-inspired Systems and Signal Processing (BIOSIGNALS2010). 428-431. Usage::

r_peaks = detectors.two_average_detector(unfiltered_ecg)

Matched Filter

FIR matched filter using template of QRS complex. Uses the Pan and Tompkins thresolding method. The ECG template is a text file where the samples are in a single column. See the templates folder on github for examples. Usage::

r_peaks = detectors.matched_filter_detector(unfiltered_ecg,template_file)

Heartrate variability analysis

The module hrv provides a large collection of heartrate variability measures which are methods of the class HRV::

HR(self, rr_samples) Calculate heart-rates from R peak samples.

NN20(self, rr_samples) Calculate NN20, the number of pairs of successive NNs that differ by more than 20 ms.

NN50(self, rr_samples) Calculate NN50, the number of pairs of successive NNs that differ by more than 50 ms.

RMSSD(self, rr_samples, normalise=False) Calculate RMSSD (root mean square of successive differences).

SDANN(self, rr_samples, average_period=5.0, normalise=False) Calculate SDANN, the standard deviation of the average RR intervals calculated over short periods.

SDNN(self, rr_samples, normalise=False) Calculate SDNN, the standard deviation of NN intervals.

SDSD(self, rr_samples) Calculate SDSD (standard deviation of successive differences), the standard deviation of the successive differences between adjacent NNs.

fAnalysis(self, rr_samples) Frequency analysis to calc self.lf, self.hf, returns the LF/HF-ratio.

pNN20(self, rr_samples) Calculate pNN20, the proportion of NN20 divided by total number of NNs.

pNN50(self, rr_samples) Calculate pNN50, the proportion of NN50 divided by total number of NNs.

For parameters and additional info use the python help function::

import hrv help(hrv)

The example hrv_time_domain_analysis.py calculates the heartrate variability in the timedomain.

Benchmarking

run_all_benchmarks.py calculates the R peak timestamps for all detectors, the true/false detections/misses and saves them in .csv files. Open the script itself or use python's help function of how to obtain the ECG data such as the MIT db.

show_stats_plots.py takes then the .csv files, displays the results of the different detectors and calculates the stats.

hrv_time_domain_analysis.py performs a timedomain analysis between sitting and a math test using the EngZee detector and the wavelet detector for comparison.

Realtime / Causal processing

Most ECG R-peak detectors won't detect the actual R-peak so the name "R-peak detector" is a misnomer. However in practise this won't play any role as only the temporal differences between R-peaks play a role. Most detectors work with a threshold which moves the detection forward in time and use causal filters which delay the detection. Only a few detectors do actually a maximum detection but even they will be most likely introducing delays as the ECG will be always filtered by causal filters. In other words most detectors cause a delay between the R peak and its detection. That delay should of course be constant so that the resulting HR and HRV is correct.

Authors

Luis Howell, [email protected]

Bernd Porr, [email protected]

citation / DOI

DOI: 10.5281/zenodo.3353396

https://doi.org/10.5281/zenodo.3353396