TABLE 1. COVID-19 detection: Current state of the art analysis and comparisons.
| References | Methodology | Data Type | Results | |
|---|---|---|---|---|
| Chen et al.4 | Features model (C model), (R model), and (CR model). | CT scans | sensitivity (0.961) specificity (0.957) accuracy (0.959) AUC* 0.986 (0.966∼1.000) |
|
| Ardakani et al.7 | AlexNet, VGG-16, VGG-19, SqueezeNet, GoogleNet, MobileNet-V2, ResNet-18, ResNet-50, ResNet-101, and Xception |
CT scans | AUC (0.994) sensitivity (100); specificity, (99.02) accuracy (99.51%). |
|
| Salman et al.8 | Trained CNN to detect COVID-10 | Chest X-rays | sensitivity (100) specificity (100) accuracy (100), PPV (100) NPV (100) |
|
| Ozturk et al.10 | DarkNet model and YOLO | Chest X-rays | sensitivity (85.35) specificity (92.18) accuracy (87.02) precision (89.96) F1-score (97.37) |
|
| Butt et al.11 | 3D CNN, ResNet, image preprocessing method based on HU Value, Noisy Bayesian Function | CT scans | AUC 0.996 (95%CI: 0.989–1.00) sensitivity (98.2) specificity (92.2). |
|
| Toğaçar et al.14 | Features were extracted using CNN model (MobileNetV2, SqueezeNet). Finally, for features selection, Social Mimic optimization is employed. | X-ray images | Accuracy (99.27) | |
| Singh et al.16 | A fusion of CNN, ANN, and ANFIS models to classify infected patients from COVID-19 |
Chest CT images | Accuracy (1.9789) F-measure (2.0928) sensitivity (1.8262), specificity (1.6827) Kappa statistics (1.9276) |
|
| Wu et al.17 | Multi-view deep learning fusion model | CT images | AUC (0.732) accuracy (0.700) sensitivity (0.730) specificity (0.615) in validation set. For test set AUC (0.819) accuracy (0.760) sensitivity (0.811) specificity (0.615) |
|
| Ucar and Korkmaz18 | Deep Bayes-SqueezeNet based rapid diagnostic system | Chest X-ray images | COR (98.26) COM (98.26) accuracy (98.26) specificity (99.13) MCC (97.39) F1-score (98.25) |
|
| Hasan et al.19 | Q-Deformed Entropy Feature Extraction (QDE), convolutional neural network (CNN) feature extractor and LSTM Neural Network Classifier. | CT Images | Accuracy (99.68) TP COVID-19 (100) TP Healthy (100) TP Pneumonia (98.90) |
|
| Loey et al.20 | Generative Adversarial Network (GAN) and Convolutional Neural Networks (AlexNet, Googlenet, Resnet18) | Chest X-ray images | Googlenet 4 classes Test accuracy (80.56) F1-score (82.32) Recall (80.56) Precision (84.17) |
|
*AUC (Area Under Curve)Common benchmarking dataset.