Table 5.
Deep learning architecture specially used in COVID-19 and compared with other architecture
| Image type | Publication | Custom deep learning architecture | Pre-train architecture or comparison of other architecture | No. of class with class name | Results (%) |
|---|---|---|---|---|---|
| X-ray | Khan el al [13] | CoroNet | VGG-19, Mobile Net, ResNet |
No of class: 3 COVID-19, Normal, Pneumonia, Bacterial |
Acc: 90.00, PR: 93.17, RE: 98, SP: 97.9, F: 95.61 |
| Gupta el al [2] | InstaCovNet-19 |
CovidAID, COVIDiagnosis-Net, CoroNet, ResNet50 + DCNN, COVID-Net, DarkCovidNet, MobileNet v2 and CovidAID |
No. of class: 3 Normal, Pneumonia, COVID-19 |
Acc: 99.08, F: 99.00, PR: 99.00 |
|
| Mahmud el al [190] | CovXNet |
DarkCovidNet, COVID-Net, VGG-19, ReNet-50/SVM, ResNet50 |
No. of class: 4 Normal, Viral pneumonia, Bacterial pneumonia, COVID-19 |
Acc: 91.70, AUC: 94.10, PR: 92.90, RE: 92.10, SP: 93.60, F: 92.60 |
|
| Madaan et al. [191] | XCOVNet |
Three dimensional-CNN classification, Inception, ResNet50 + SVM, CNN + ResNet + Inception |
No. of class: 2 COVID-19, Normal |
Acc: 98.71, PR: 97.95, RE: 69.89, F: 82.28 |
|
| Zhang el al. [123] | COVID19Xray-Net | ResNet34 |
No. of class: 3 Normal, Pneumonia, COVID-19 |
Acc: 91.92, AUC: 98.50, SP: 90.00, SN: 97.00 |
|
| Chowdhury el al. [192] | PDCOVID-Net | VGG16, ResNet50, InceptionV3, DenseNet121 |
No. of class: 3 COVID-19, Normal, Viral pneumonia |
Acc: 96.58, PR: 96.58, RE: 96.59, F: 96.58 |
|
| Ucar and D. Korkmaz [193] | COVIDiagnosis-Net | DenseNet, Tailored CNN, Capsule Networks, ResNet50, Sgdm-SqueezeNet |
No. of class: 3 COVID-19, Normal, Pneumonia |
Acc: 98.26, CR: 98.30, CM: 98.30, SP: 99.10, F: 98.30, MCC: 97.4 |
|
| Ozturk el al. [194] | DarkCovid Net |
COVID-Net, COVIDX-Net, ResNet50, VGG-19, M-Inception, DRE-Net |
No. of class: 2 COVID-19, non-COVID-19 |
Acc: 95.13, SN: 85.35, SP: 92.18, PR: 89.96, F: 87.37 |
|
| Ambati el al. [195] | AC-CovidNet |
DarkCovidNet Covid-Caps, CovXnet COVID-Net, CovidAID, |
No. of class: 3 COVID, Pneumonia, Normal |
Acc: 96.66, SN: 96.66 |
|
| Hanafi et al. [196] | Cae-covidx |
VGG16, Custom CNN |
No. of class: 2 Normal, COVID |
Acc: 98.00 | |
| X-ray | Hussain et al. [197] | CoroDet |
COVID-Net, COVIDX-Net, Inception-ResNetV2, ResNet152 |
No. of class: 4 COVID-19, Normal, Viral pneumonia, Bacterial pneumonia |
Acc: 91.2, SN: 91.76, SP: 93.48, PR: 92.04, RE: 91.9, F: 90.04 |
| Elbishlawi et al. [198] | Corona-Net |
VGG-19, ResNet50, COVID-Net |
No. of class: 3 COVID-19, Pneumonia, Normal |
Acc: 100.00, F: 99.80 |
|
| Haghanifar et al. [199] | COVID-CXNET |
ShuffleNetV2, InceptionV3, COVID-Net, DenseNet |
No. of class: 3 Normal, COVID-19, Pneumonia, non-COVID-19 |
Acc: 87.21, F: 92.21 |
|
| Khan et al. [200] | CovidMulti-Net | Pre-train Model |
No. of class: 4 Normal, COVID-19, Viral pneumonia, Bacterial pneumonia |
Acc: 98.40, PR: 83.00, RE: 100.00, F: 91.0 |
|
| Wang et al. [201] | COVID‑Net |
VGG19, ResNet50 |
No. of class: 2 COVID-19, non-COVID-19 |
Acc: 93.30, SN: 91.00 |
|
| R. K. Singh et al. [202] | COVIDScreen |
Covid-net, Coronet, COVID-CXNet, CovidAID |
No. of class: 3 Normal, COVID-19, Pneumonia |
Acc: 98.67, PR:100.00, RE: 100.00, F: 100.00 |
|
| Das et al. [203] | CoviLearn |
CovidNet, VGG-19, ResNet50, DarkNet |
No. of class: 2 Normal, COVID-19 |
Acc: 98.98, SN: 100.00, SP: 98.00, AUC: 99.00 |
|
| Umer et al. [204] | COVINet |
VGG16, AlexNet |
No. of class: 4 Normal, COVID-19, Virus pneumonia, Bacterial pneumonia |
Acc: 84.76, PR: 89.29, RE: 98.99, F: 93.89 |
|
| Hertel et al. [205] | COV-SNET |
Covid-cxnet, ChestX-Ray8, Deepcovid-xr, Covid-net |
No. of class: 3 COVID-19, Normal, Pneumonia |
Acc: 83.30, SN: 95.00, SP: 85.00, F: 86.00 |
|
| Ouchicha et al. [206] | CVDNet |
MobileNetV2, VGG19, InceptionV3, DenseNet201, InceptionResNetV2, ResNetV2, Xception |
No. of class: 3 COVID-19, Normal, Pneumonia |
Acc: 94.00, SP: 96.00, SN: 90.00, AUC: 96.00 |
|
| Minaee et al. [207] | Deep-COVID |
ResNet18, ResNet50, SqueezeNet, DenseNet121 |
No. of class: 2 COVID-19, non-COVID-19 |
SN: 98.00, SP: 92.00 |
|
| Quan et al. [208] | DenseCapsNet |
DenseNet, ResNet50, CapsNet |
No. of class: 3 Normal, Pneumonia, COVID-19 |
Acc: 90.70, F: 90.90 |
|
| Cheng et al. [209] | DPN-SENet |
ResNet, DenseNet, DPNNet, VGG16, Inceptionv4 |
No. of class: 4 COVID-19, Normal, Bacterial pneumonia, Viral pneumonia |
Acc: 93.00, RE: 98.00, PR: 97.00, F: 98.00 |
|
| Chowdhury et al. [210] | ECOVNet |
COVID-Net, EfficientNet-B3, DarkCovidNet, CoroNet |
No. of class 3 COVID-19, Normal, Pneumonia |
Acc: 95.00, PR: 90.29, RE: 93.00, F: 91.63 |
|
| X-ray | Tang et al. [211] | EDL-COVID |
ResNet50, Inception V3, Inception-ResNetV2 |
No. of class 3 COVID-19, Normal, Pneumonia |
Acc: 95.0, SN: 96.0 |
| Agrawal et al. [212] | FocusCovid |
DarkCovidNet, nCOVNet, Capsnet, COVID-Net, ResNet18, VGG-19 |
No. of class 3 COVID-19, Normal, Pneumonia |
Acc: 95.20, PR: 95.0, SN: 95.20, F: 95.20 |
|
| Monshi et al. [213] | CovidXrayNet |
VGG-16, VGG-19 ResNet18,ResNet34 ResNet50, EfficientNet-B0 |
No. of class 3 COVID-19, Normal, Pneumonia |
Acc: 95.82, AUC: 99.29, F: 96.16, PR: 96.93, RE: 95.43 |
|
| Toraman et al. [214] | Convolutional CapsNet | DarkCovidNet, VGG19, MobileNetV2 DeCovNet and DRE-Net |
No. of class 3 COVID-19, Pneumonia, No finding |
Acc: 97.24, SP: 98.00, |
|
| Hemalatha et al. [215] | FractalCovNet |
ResNet5, Xception, InceptionResNetV2, VGG-16, Den- seNet |
No. of class 2 COVID-19, Others |
Acc: 99.00 PR: 99.00 RE: 87.00 F: 92.00 |
|
| CT | Ghaderzad et al. [216] | NASNet |
SqueezeNet, ShuffleNet, VGG-19, Modified VGG-19, COVID-CT-Net |
No of class 2 COVID-19, non-COVID-19 |
Acc: 99.60, SP: 98.60, SN: 99.90, |
| Ibrahim et al. [217] | Deep-chest |
VGG19-CNN, ResNet152V2, ResNet152V2 + Gated Recurrent Unit (GRU), ResNet152V2 + Bidirectional GRU (Bi-GRU) |
No. of class: 4 Covid-19, Normal, Pneumonia, Lung cancer |
Acc: 98.05, RE: 98.05, SP: 98.43, F: 98.24, AUC: 99.66 |
|
| Shah el al [218] | CTnet-10 |
ResNet50, InceptionV3, DenseNet- 169, VGG-16, VGG-19 |
No. of class: 2 COVID-19, non-COVID-19 |
Acc: 94.52 | |
| Bhansali et al. [219] | CoronaNet | Pure CNN based |
No. of class: 2 COVID-19, Normal |
Acc: 92.14, PR: 91.07, RE: 89.38, F: 87.35 |
|
| Wu et al. [220] | COVID-AL | Not reported |
No. of class: 3 COVID-19, Pneumonia, Common pneumonia, Normal |
Acc: 86.6, R-AUC: 96.62, ROC-AUC: 96.8 |
|
| Javaheri et al. [221] | CovidCTNet |
EfficientNet B4, ResNet18, Resnet50, 3D Resnet-18 |
No. of class: 2 COVID-19, non-Covid-19 |
Acc: 95.00, PR: 88.00, RE: 74.00, F: 80.00 |
|
| X-ray/CT | Afshar el al [222] | COVID-CAPS | ImageNet |
No. of class: 4 Normal, Bacterial pneumonia, Viral pneumonia, COVID-19, non-COVID-19 |
Acc: 83.50, AUC: 97.00, SN: 90.00, SP: 95.80 |
Acc accuracy, PR precision, RE recall, SP specificity, F F-measure, SN sensitivity, MCC Matthew correlation coefficient, CR correctness, CM completeness, projection expansion projection extension (PEPX), FC fully connected, BN batch normalization