Table 10.
Comparison to traditional AI approaches.
| Approach | Sensitivity (%) | Specificity (%) | Precision(%) | Accuracy (%) |
|---|---|---|---|---|
| Multiscale AM-FM (Murray et al., 2010) | 94.08 | 93.64 | 91.91 | 93.83 |
| ARF (Nayak et al., 2016) | 96.23 ± 1.18 | 96.32 ± 1.48 | N/A | 96.28 ± 1.25 |
| BWT-LR (Wang et al., 2016) | 97.12 ± 0.14 | 98.25 ± 0.16 | N/A | 97.76 ± 0.10 |
| 4-level HWT (Wu and Lopez, 2017) | N/A | N/A | N/A | 87.65 ± 1.79 |
| MBD (Zhang et al., 2017) | 97.78 ± 1.29 | 97.82 ± 1.60 | N/A | 97.80 ± 1.40 |
| CNN-DO-BN-SP (Ours) | 98.77 ± 0.35 | 98.76 ± 0.58 | 98.75 ± 0.58 | 98.77 ± 0.39 |