Table 4.
The comparison of the proposed approach compared to existing methodologies.
| Authors | Methods | Dataset | Classes # | Acc (%) | Se (%) | Sp (%) |
|---|---|---|---|---|---|---|
| Ozturk et al. [21] | DarkCovidNet | Public | 3 | 87.02 | 92.18 | 89.96 |
| Wang et al. [9] | COVID-Net | Public | 3 | 92.64 | 91.37 | 95.76 |
| Apostolopoulos et al. [19] | The pre-trained CNNs | Public | 3 | 96.78 | 98.66 | 96.46 |
| Ucar and Korkmaz [41] | COVIDiagnosis-Net | Public | 3 | 98.26 | 98.33 | 99.10 |
| Nour et al. [42] | Deep CNN, SVM | Public | 3 | 98.97 | 89.39 | 99.75 |
| Turkoglu [43] | AlexNet, Feature Selection, SVM | Public | 3 | 99.18 | 99.13 | 99.21 |
| Togacar et al. [44] | Deep features, SqueezeNet, SVM | Public | 3 | 99.27 | 98.33 | 99.69 |
| Demir et al. [27] | DeepCov19Net | Public | 3 | 99.75 | 99.33 | 99.79 |
| Demir [24] | DeepCoroNet | Public | 3 | 100.00 | 100.00 | 100.00 |
| Ismael and Sengur [25] | ResNet50 Features + SVM | Public | 2 | 94.74 | 91.00 | 98.89 |
| Muralidharan et al. [26] | FB2DEWT + CNN | Public | 3 | 96.00 | 96.00 | 96.00 |
| Proposed Method | Processed images, ACL model | Public | 3 | 100.00 | 100.00 | 100.00 |