Table 5.
Comparative performance analysis of CIDICXR-NET50 against other leading techniques.
| Model | Dataset | Methodology | Dataset size | Classification | Accuracy % |
|---|---|---|---|---|---|
| Sethy et al. (30) | CXR Images | TL ResNet50 + SVM | 381 | Multi-class | 95.33 |
| Pathak et al. (37) | CT Images | TL ResNet50 | 852 | Binary-class | 93.01 |
| Ozturk et al. (34) | CXR Images | TL DarkCovidNet | 1,127 | Multi-class | 87.02 |
| Chowdhury et al. (49) | CXR Images | TL DenseNet-201 | 3,487 | Binary-class | 99.70 |
| Chowdhury et al. (49) | CXR Images | TL DenseNet-201 | 3,487 | Multi-class | 97.94 |
| Apostolopoulos & Mpesiana (29) | CXR Images | TL MobileNetV2 | 1,442 | Multi-class | 94.72 |
| Wang et al. (28) | CXR Images | TL COVID-Net | 13,975 | Multi-class | 93.3 |
| Hemdan et al. (31) | CXR Images | TL COVIDX-Net | 53 | Binary-class | 90 |
| Jain et al. (Phase I) (35) | CXR Images | TL ResNet-50 | 1832 | Multi-class | 93 |
| Jain et al. (Phase II) (35) | CXR Images | TL ResNet101 | 1832 | Binary-class | 97.78 |
| Manokaran et al. (32) | CXR Images | TL DenseNet-201 | 8,644 | Multi-class | 92.19 |
| Chakraborty et al. (33) | CXR Images | TL VGG-19 | 3,797 | Multi-class | 97.11 |
| CIDICXR-Net50 (Proposed) | CXR Images | TL ResNet-50 | 3,923 | Multi-class | 99.11 |