Table 1.
Summary of techniques available in literature for COVID19 screening
| Ref. | Technique | Key findings | Dataset |
|---|---|---|---|
| [11] | Transfer learning on ResNet-50 (CNN model) | Achieved 10-fold cross-validation accuracy of 93.01% on 109 test images. | 413 COVID-19 (+) images and 439 images of normal or pneumonia infected patients. [12] |
| [13] | Resnet50 and VGG16 (deep learning) | COVID-19 positive cases and pneumonia cases of X-ray modalities are classified with an accuracy of 89.2%. | There were 135 COVID-19 cases obtained from JP Cohen [14], radiopedia and SIRM [15]. |
| [16] | Two-step transfer learning model | Two-step transfer learning pipeline based on the deep neural network framework COVID19XrayNet. The approach achieved a maximum accuracy of 91.4%. | The study uses 189 COVID-19 Chest X-Ray images (131 train and 41 test). [14] |
| [17] | EfficientNet | An accuracy of 93.9%, a sensitivity of 96.8%, and positivity prediction of 100% are obtained on 231 test X-ray images (COVID19-positive cases—31, pneumonia—100, and normal cases—100). | The model is trained on 13569 X-ray images (COVID19 positive cases-152, pneumonia-5421, and normal cases-7966) [18] |
| [19] | Pre-trained CheXNet and DenseNet | An accuracy of 90.5% and a sensitivity of 100% are achieved on 654 test X-ray images (COVID19-positive cases—30, pneumonia—390, and normal cases—234). | The model is trained on 5323 chest X-ray images (COVID19 positive cases-115, pneumonia-3867, and normal cases - 1341) [14]. |
| [20] | Domain extension transfer learning (DETL) with gradient class activation map (Grad-CAM) | Fivefold cross-validation accuracy of 90.13% and test set accuracy of 95.3% are obtained on the proposed X-ray dataset. | A total of 305 COVID-19 X-Ray images were used in the study. [14] |
| [21] | ResNet, Inception, and GoogleNet | The classification of COVID-19 positive cases based on X-ray modality is done. The approach achieved 98% of accuracy with VGG19, 95% with Resnet50, and 96% with InceptionV3. | A dataset of nearly 100 subject, among them 50 x-ray images subjects were tested positive with COVID-19 and 50 x-ray images of normal subjects. [14, 15] |
| [22] | ResNet18, ResNet50, SqueezeNet, and DenseNet-121 | The model achieved a specificity of 90% and sensitivity of 96.5% on testing data of 3000 chest X-rays (COVID and non-COVID patients). | The model is trained on 5000 chest x-rays dataset [14] |
| [23] | Joint classification and segmentation (JCS) | Classification is done with a specificity of 93% and a sensitivity of 95%. A dice score of 78.3% is obtained for the segmentation task. | JCS system is implemented on 400 COVID-19 patients (144,167 images) and 350 Non-COVID patients. [14, 24, 25] |
| [26] | Pruned efficient net-based model on chest CT scans and X-rays | Classification into two binary classes, i.e., COVID and non-COVID. The highest accuracy of 85.22% is achieved with the ResNet50 pre-trained CNN model. | The CNN based pre-trained models are trained on 596 chest CT scans. [14, 15] |
| [27] | Detail-oriented capsule networks (DECAPS) +Peekaboo (patch crop and drop strategy) | An accuracy of 87.6%, recall of 91.5%, precision of 84.3%, and AUC of 96.1 are achieved for binary classification (COVID-19 and non-COVID) of chest CT scan. | Uses a total of 746 chest CT images - COVID-19 (349 images) and non-COVID-19 (347 images). [15] |
| [28] | Transfer learning on Xception net | For binary classification of chest CT scan of COVID and non-COVID dataset, the model achieved a sensitivity of 96.1%, the specificity of 93.4%, and AUC of 0.92. | It contains three classes as COVID-19 (+), pneumonia (+) but COVID-19 (-) with 504 images. [25] |
| [29] | Multi-objective differential evolution (MODE) deep learning | In comparison to authentic CNN models, the performance parameters of MODE outperforms by 2.09% of F-measure, 1.82% of sensitivity, 1.68% of specificity, and 1.927% of Kappa statistics. | A study of 73 patients with 205 COVID positive images. [30] |