Table 5.
Benchmarking table-showing comparison of proposed and existing classification (solo) models.
| Author and Year | Dataset—Chest X-ray (COVID-19 Images + Other Images) | Technique | Accuracy | AUC |
|---|---|---|---|---|
| Nayak et al. (2020) [29] | GitHub (203 + 203) |
ResNet-34 | 2 class-98.33% | 2 class-0.98 |
| Choudhury et al. (2020) [60] | COVID-19 Radiography database (Kaggle) (423 + 3064) |
CheXNet | 3 class-97.74% | NA |
| Jain et al. (2020) [28] | Kaggle (490 + 5942) |
Xception | 3 class-97.97% | NA |
| Nikolaou et al. (2021) [68] | COVID-19 Radiography database (Kaggle) (3616 + 11,537) |
EfficientNetB0 | 2 class-95% 3 class-93% |
NA |
| Yang et al. (2021) [83] | COVID-19 Radiography database (Kaggle) (3616 + 4845) |
VGG16 | 2 class-98% 3 class-97% |
NA |
| Khan et al. (2020) [26] | GitHub (284 + 967) |
Coronet (novel CNN) | 3 class-95% | NA |
| Hussain et al. (2020) [27] | COVID-R dataset (500 + 1600) |
CoroDet (novel CNN) | 2 class-99.1% 3 class-94.2% 4class-91.2% |
NA |
| Aslan et al. (2020) [84] | COVID-19 Radiography database (Kaggle) (219 + 2686) |
mAlexNet + BiLSTM (Bidirectional long short term memory) | 3 class-98.7% | NA |
| Timemy et al. (2021) [85] | GitHub (435 + 1751) |
ResNet-50 + ESD (Ensemble Subspace Discriminant) | 5 class- 91.6% | NA |
| Khan et al. (2022) [86] | COVID-19 Radiography database (Kaggle) (3616 + 17,449) |
EfficientNetB | 4 class-96.13% | NA |
| Nillmani et al. (2022) [69] | COVID-19 Radiography database (Kaggle) (3611 + 13,833) |
VGG16, NASNetMobile, DenseNet201 | 2 class-99.84% 3 class-96.63 5 class-92.70 |
2 class-1.0 3 class-0.97 5 class-0.92 |
| Proposed | COVID-19 Radiography database (Kaggle) (3611 + 9849) |
Xception | 5 class-97.45% | 0.998 |