Research Article
SS-SWT and SI-CNN: An Atrial Fibrillation Detection Framework for Time-Frequency ECG Signal
Table 4
Comparison of the performances of AF classification algorithms.
| Algorithm | Signal lengths | Methodology | Se (%) | Sp (%) | Acc (%) |
| Tateno and Glass [37] | 50 s | RR interval irregular | 94.4 | 97.2 | — | Dash et al. [9] | 128 beats | RR interval irregular | 94.4 | 95.1 | — | Babaeizadeh et al. [38] | >60 s | RR interval irregular | 92 | 95.5 | — | Huang et al. [39] | 101 beats | RR interval irregular | 96.1 | 98.1 | — | Asgari et al. [20] | 9.8 s | SWT + SVM | 97 | 97.1 | 97.1 | Ladavich and Ghoraani [40] | 7 beats | P-wave absence (PWA) | 98.09 | 91.66 | 93.12 | García et al. [21] | 7 beats | SWT | 91.21 | 94.63 | 93.32 | He et al. [33] | 5 beats | SWT | 99.41 | 98.91 | 99.23 | Xia et al. [32] | 5 s | SFWT | 98.34 | 98.24 | 98.29 | Xia et al. [32] | 5 s | SWT | 98.6 | 97.17 | 97.74 | Proposed framework | 5 s | SWT | 99.03 | 99.35 | 99.23 |
|
|