Table 5.
Performance comparison of the proposed 1D deep CNN model with the previous studies for the per-minute apnea detection.
| Reference | Methods | Sen (%) | Spe (%) | Acc (%) | AUC |
|---|---|---|---|---|---|
| Feature-Learning-Based Methods | |||||
| Our Study | The proposed 1D Deep CNN Model | 81.1 | 92.0 | 87.9 | 0.94 |
| Singh and Majumder [15] | Pre-trained AlexNet CNN + Decision Fusion |
90.0 | 83.8 | 86.2 | 0.88 |
| Wang et al. [16] | LeNet-5 CNN | 83.1 | 90.3 | 87.6 | 0.95 |
| Li et al. [17] | Auto-encoder + Decision Fusion | 88.9 | 82.1 | 84.7 | 0.87 |
| Feature-Engineering-Based Methods | |||||
| Sharma and Sharma [12] | Feature Engineering + LS-SVM | 79.5 | 88.4 | 83.8 | 0.83 |
| Song et al. [13] | Feature Engineering + HMM-SVM | 82.6 | 88.4 | 86.2 | 0.94 |
| Varon et al. [14] | Feature Engineering + LS-SVM | 84.7 | 84.7 | 84.7 | 0.88 |