Table 2.
Remaining summary of DL-based Covid-19 X-Ray, CT diagnosis systems
| Author | Type | Training Model | Resol-ution | Total Images |
|---|---|---|---|---|
| Wang et al. [95] | Pre-trained (XRay) | RestNet50, ResNet101, ResNet152 | 18,567 | |
| Arellan,Ramos [53] | Pre-trained (XRay) | DenseNet121 | 38 | |
| Minaee et al. [56] | Pre-trained (XRay) | Deep-COVID (ResNet18,ResNet50, SqueezeNet, DenseNet-121) | 5420 | |
| Demir [99] | Custom (XRay) | DeepCoroNet | 100 × 100 | 1061 |
| Sheykhivand et al. [52] | Pre-trained (XRay) | Inception V4 | 224 × 224 | 11,383 |
| Mishra et al. [55] | Pre-trained (XRay) | CovAI-Net (Inception, DenseNet, Xception) | 224 × 224 | 1878 |
| Sakib et al. [100] | Custom (XRay) | DL-CRC | 2905 | |
| Tang et al. [44] | Custom (XRay) | EDL-COVID | 15,477 | |
| Saha et al. [94] | Pre-trained (XRay) | EMCNet(AlexNet, VGG 16, Inception, and ResNet-50) | 224 × 224 | 4600 |
| Gupta et al. [54] | Hybrid(XRay) | InstaCovNet-19 ( InceptionV3,NasNet, Xception,Mobile NetV2, ResNet101) | 224 × 224 | 3047 |
| Vaid et al. [134] | Pre-trained (XRay) | VGG19 | 224 × 224 | 545 |
| Bhosale et al. [47] | Custom (XRay) | LDC-Net (IoT based) | 1024 × 1024 | 10,800 |
| Serener et al. [111] | Pre-trained (CT) | ResNet-50, ResNet-18, MobileNetV2, VGG,AlexNet,SqueezeNet, DenseNet121 | 224 × 224 | 1005 |
| Voulodimos et al. [96] | Pre-trained (CT) | FCN-8, U-Net | 630 × 630 | 939 |
| Chen et al. [1] | Pre-trained (CT) | ResNet50, Unet + + | 512 × 512 | 80,030 |
| Wu et al. [50] | Pre-trained (CT) | ResNet50 | 256 × 256 | 495 |
| Shah et al. [108] | Custom, Pre-trained (CT) | CTNet-10,DenseNet 169, VGG16/19, ResNet50,InceptionV3, | 128 × 128 to 224 × 224 | 812 |
| Khan et al. [102] | Hybrid (CT) | H3DNN(3DResNet, C3D, 3D DenseNet, I3D, LRCN) | 224 × 224 | 880 |
| Author | Classes | Partition | Performance | Data Source | Time |
|---|---|---|---|---|---|
| Wang et al. [95] | 8851(Normal), 9576(Pneumonia), 140(Covid19) | NA | Accuracy:96.1% | [72, 80] | NA |
| Arellan,Ramos [53] | 19(Covid + Ve), 19(Covid-Ve) | NA | Accuracy:99%, Recall:89%,Precision:91%,Fscore: 89% | [73, 82] | NA |
| Minaee et al. [56] | 5000(Normal), 420(Covid19) | Random | Sensitivity:98%, Specificity:92.9% | [72, 90] | NA |
| Demir [99] | 361(Covid19), 200(Normal), 500(Pneumonia) | Training:80%, Testing:20% | Accuracy:100%, Sensitivity:100%, Specificity:100% | [71, 82, 113], | Train: 53 M 22 s |
| Sheykhivand et al. [52] | 2923(Healthy), 2842(Covid19), 2778(Bacterial), 2840(Viral) | Training:70%, Testing:10%, Validation:20% | Accuracy:99.5%, Sensitivity100%, Specificity:99.02% | [71–74] | Test: 3 s |
| Mishra et al. [55] | 570(Pneumonia), 630(Non-pneumonia), 369(Covid19 +), 309(Covid19 -) | Random | Accuracry:98.31%, Precision:100%, Sensitivity:96.74%, Specificity:100%, F1-Score:98.34% | [77, 80], ESR | NA |
| Sakib et al. [100] | 219 (Covid19 +), 1341(Normal), 1345(pneumonia) | fivefold cross validation | Accuracy:93.94%, AUC:95.25% | [71, 72, 82] | NA |
| Tang et al. [44] | 6053(Pneumonia), 8851(Normal), 573(Covid19) | Accuracy:95%, Sensitivity:96.0%, PPV:94.1% | [120] | 1 s Exec | |
| Saha et al. [94] | 2300(Covid19), 2300(Normal) | Training:70%, Testing:10%, Validation:20% | Accuracy:98.91%, Precision:100%, Recall:97.82%, F1-score:98.89% | [71, 72, 75, 77, 80, 82, 133], | NA |
| Gupta et al. [54] | 1345(Pneumonia), 1341(Normal), 361(Covid19) | Training:80%, Testing:20% | Accuracy:99.53%, Precision:100%, Recall:99%, F1-Score:99% | [73, 82, 90] | NA |
| Vaid et al. [134] | 181(Covid19), 364(NoFinding) | Training:80%, Testing:20%, Validation:20% | Accuracy:96.3% | [72, 82] | NA |
| Bhosale et al. [47] | Covid-19, other 8 lung diseases | Train:76%, Test:12%, Val:12% | Acc:96.3%,Recall: 96.78%,Fscore:96.77%, AUC:98.18% | [82] and other 4 datasets | 0.136 s |
| Serener et al. [111] | 397(Mycoplasma Pneumonia), 145 (ViralPneumonia), 463(Covid19) | Random | Accuracy:89%, Sensitivity:98%, Specificity:86%, AUC:95% | [81] | NA |
| Voulodimos et al. [96] | 447(Covid-negative), 492(Covid-positive) | Training:85%, Validation:15% | Accuracy:99%, Recall:89%, Precision:91%, F1-Score: 89% | [80] | NA |
| Chen et al. [1] | 49,089(Covid19), 30,941(Normal) | Random | Accuracy:96%, Sensitivity:98%, Specificity:94%, PPV:94.23%, NPV:97.92% | Renmin Wuhan Univ., Qianjiang Hospital, China | NA |
| Wu et al. [50] | 368(Covid19), 127(other pneumonia) | Training:80%, Testring:10%, Validation:10% | Accuracy:76%, AUC:81.9%, Sensitivity:81.1%, Specificity:61.5% | China Medical Univ., BYH in China | < 5 s |
| Shah et al. [108] | 349(Covid19 confirmed), 463(nonCovid19) | Training:80%, Testring:10%, Validation:10% | CTNet Accuracy:82.1%, VGG19 Accuracy: 94.52% | [76] | Train: 130 s, Test:0.9, Exe: 0.01233 s |
| Khan et al. [102] | 417(Covid19), 463(Non-Covid19) | NA | Accuracy:85% | [82, 86] | NA |
NA indicates the corresponding author did not disclose the parameter value