Table 3.
Selected arrhythmia detection studies using HR and ECG. Database (DB) were: MIT-BIH Atrial Fibrillation Database (afdb), MIT-BIH Arrhythmia Database (mitdb), MIT-BIH Malignant Ventricular Arrhythmia Database (vfdb), Creighton University Ventricular Tachyarrhythmia Database (cudb), MIT-BIH Normal Sinus Rhythm Database (nsrdb), MIT-BIH Long Term Database (ltafdb), European ST-T Database (edb) and ecgdb. Hospital data come from non-publicly accessible databases.
| Author Year | Method | Data | Performance | ||||
|---|---|---|---|---|---|---|---|
| Type | DB | Rhythm | ACC | SPE | SEN | ||
| Faust et al. 2020 [43] | Detrending, ResNet | HR | ecgdb | AF Atrial Flutter (AFL) Normal Sinus Rhythm (NSR) | 99.98 | 100.00 | 99.94 |
| Ivanovic et al., 2019 [44] | CNN, LSTM | HR | Hospital | NSR, AF AFL | 88 | 87.09 | |
| Fujita and Cimr, 2019 [45] | CNN with normalization | ECG | afdb, mitdb, vfdb | AF, AFL, VFDB, NSR | 98.45 | 99.87 | 99.27 |
| Faust et al., 2018 [14] | LSTM | HR | afdb | AF NSR | 98.39 | 98.32 | 98.51 |
| Acharya et al., 2017 [46] | CNN with Z-score | ECG | afdb, mitdb, vfdb | AF, AFL, VFIB, NSR | 92.50 | 98.09 | 93.13 |
| Henzel et al., 2017 [47] | Statistical features with generalized linear model | HR | afdb | AF NSR | 93 | 95 | 90 |
| Desai et al., 2016 [48] | RQAwith decision tree, random forest, rotation forest | ECG | afdb, mitdb, vfdb | AF, AFL, VFIB, NSR | 98.37 | ||
| Acharya et al., 2016 [49] | Thirteen nonlinear features with ANOVA with KNN and DT | ECG | afdb, mitdb, vfdb | AF, AFL, VFIB, NSR | 97.78 | 99.76 | 98.82 |
| Hamed and Owis, 2016 [50] | DWT, PCA and SVM | ECG | afdb | AF, AFL, NSR | 98.43 | 96.89 | 98.96 |
| Xia et al., 2018 [51] | STFT/SWTwith CNN | ECG | afdb | AF | 98.63 | 98.79 | 97.87 |
| Petrėnas et al., 2015 [52] | Median filter with threshold | HR | nsrdb, afdb | AF NSR | 98.3 | 97.1 | |
| Zhou et al., 2014 [53] | Median filter and Shannon entropy with threshold | HR | ltafdb, afdb, nsrdb | AF NSR | 96.05 | 95.07 | 96.72 |
| Muthuchudar and Baboo, 2013 [54] | UWT NN | ECG | afdb | AF, VFIB, NSR | 96 | ||
| Yuanet al., 2016 [55] | Unsupervised autoencoder NN Softmax regression | ECG | afdb, nsrdb, ltdb, hospital | AF | 98.18 | 98.22 | 98.11 |
| Pudukotai Dinakarrao and Jantsch, 2018 [56] | Daubechies-6 with counters Anomaly detector | ECG | mitdb | AF, VFIB | 99.19 | 98.25 | 78.70 |
| Salem et al., 2018 [57] | Spectrogram with CNN | ECG | afdb, nsrdb, vfdb and edb | AF, AFL VFIB NSR | 97.23 | ||