Skip to main content
. 2022 Sep 16:1–53. Online ahead of print. doi: 10.1007/s11063-022-11023-0

Table 4.

Summary report of work done on experimental setup in the studied papers

Author Model Layer Kernel Size Pool size Stride, Batch Size Image Size
Panwar et al. [57] VGG-19 Conv:16, Maxpool:5, FCNN:3, 3 × 3 2 × 2 Batch size:16, Stride:2 512 × 512
Nath et al. [34] CNN Conv:6, BN:6, Pool:4 3 × 3 2 × 2 Stride:2 256 × 256
Kassani et al. [97] MobileNet, DenseNet, Xception, ResNet, InceptionV3, InceptionRes-NetV2, VGGNet, NASNet Standard layer 331 × 331NASNetLarge, 224 × 224NASNetMobile, 600 × 450
Hussain et al. [93] CoroDet Conv:9, pool:9, dense:2, ft:1,LR:1 Batch size:10 256 × 256
Gilanie et al. [135] CNN Conv:8, pool:2, fc:4 3 × 3 2 × 2 Batch size:128 512 × 512
Silva et al. [106] EfficientNet Conv:3, Mbconv:7 3 × 3 Batch size:32 104 × 153, 484 × 416
Turkoglu[49] MKs-ELM-DNN DenseNet201:3 STD layer block + ELM Batch size:100 224 × 224
Horry et al. [101] VGG16/19, Resnet50, Inception V3, Xception, InceptionResNet, DenseNet, NASNetLarge Standard layer Batch size:2 and 16 224 × 224( VGG), 299 × 299 (Inception)
Dutta et al. [112] CNN, Inception V3 Batch size:32
Mertyüz et al. [51] VGG-16, ResNet, GoogleNet, Standard layer 3 × 3 1 × 1 Batch size:8
Ko et al. [38] FCONet, VGG16, ResNet-50, Inception-v3, Xception Standard layer, fc:2 Batch size:32 256 × 256
Alazab et al. [67] VGG16 Standard layer 3 × 3 1 × 1 Batch size:25 224 × 224
Sharma et at [68] VGG, MobileNet, Xception, DenseNet, InceptionResNet Standard layer Batch size:8 224 × 224
Apostolopoulos et al. [66] VGG19, MobileNetv2, Inception, Xception, Inception-ResNetv2 Standard layer Batch size:64 200 × 266
Wu et al. [115] Covid-AL 3 × 3 1 × 1 Batch size:30 352 × 320
Al-Waisy et al. [39] COVID-CheXNet (HRNet, ResNet34) 7 × 7 3 × 3 Batch size:100
Haghanifar et al. [65] COVID-CXNet 224 × 224
Sarker et al. [69] COVID-DenseNet DenseB:4, TraLayer:3 Batch size:5 224 × 224
Wang et al. [70] COVID-Net 7 × 7 to 1 × 1 batch size:64
Mangal et al. [63] CovidAID(ChexNet, Covid-Net) Batch size:16 224 × 224
Javaheri et al. [40] CovidCTNet(BCDU-Net, U Net) 3DConv:10 3DMaxPool:5, Dense:2 128 × 128
Elkorany et al. [41] COVIDetection-Net (ShuffleNet, SqueezeNet) 300 × 300
Tabik et al. [46] COVIDSDNet Batch size:16
Ucar et al. [103] COVIDiagnosis-Net (Bayes-SqueezeNet) 3 × 3 1 × 1 Batch size:32 227 × 227
Hemdan et al. [42] COVIDX-Net (VGG19, DenseNet201, InceptionV3,ResNetV2, InceptionResNetV2, Xception, MobileNetV2) Standard layer Batch size:7 224 × 224
Kedia et al. [48] CoVNet-19 (DenseNet121, VGG16) Dense layer: 32 Batch size:32 224 × 224
Ouchicha et al. [45] CVDNet Conv:9, max pool:9, concat:1, ftn:1,fc:3 5 × 5, 2 × 2 Batch size:8 256 × 256
Javor et al. [91] ResNet50 Standard Layer Batch size:32 448 × 448
Ismael et al. [98] ResNet18, ResNet50, ResNet101, VGG16, VGG19 Conv:5, ReLU:5, BN:5 3 × 3 1 × 1 224 × 224
Rohila et al. [124] ReCOV-101 (ResNet50, ResNet101,DenseNet169, DenseNet201) Conv2D:23, pool:1 3 × 3 2 × 2 224 × 224
Padma et al. [35] 2DCNN 2 × 2
Jain et al. [64] Inception V3, Xception, ResNeXt Standard Layer 128 × 128
Anwar et al. [107] EfficientNet B4 Standard Layer Batch size:16 348 × 348
Sethi et al. [104] Inception V3, ResNet50, MobileNet, Xception Standard Layer Batch size:32
Ying et al. [43] DRE-Net (ResNet50) Batch size:15 512 × 512
Jiang et al. [36] VGG, ResNet, Inception-v3, DenseNet, InceptionResNetv2, Standard Layer Batch size:4 512 × 512
Yang et al. [92] DenseNet DenseBlock:4, pool:1, linear:1 Batch size:32
Serener et al. [111] ResNet50, ResNet18, MobileNetV2, VGG, SqueezeNet, AlexNet, DenseNet121 224 × 224
Basu et al. [28] DETL(AlexNet, VGGNet, ResNet50) Alex:8, Vgg:16, Res:50 11 × 11 3 × 3
Wang et al. [95] RestNet50, ResNet101, ResNet152 Batch size:64
Arellano, Ramos [53] DenseNet121 Standard Layer
Voulodimos et al. [96] FCN-8, U-Net 3 × 3 1 × 1 630 × 630
Chen et al. [1] ResNet50, Unet +  +  512 × 512
Wu et al. [50] ResNet50 Standard Layer Batch size:4 256 × 256
Minaee et al. [56] Deep-COVID (ResNet 18, ResNet50, Squeeze Net, DenseNet-121) 3 × 3 1 × 1 Batch size:20 224 × 224
Demir[99] DeepCoroNet 11 layers Batch size:6 100 × 100
Perumal et al. [23] Resnet50, VGG16, InceptionV3 Standard layer 3 × 3 2 × 2 Stride:1, Batch size:250 226 × 226
Sheykhivand et al. [52] Inception V4 Conv:4,pool:4, lstm:2: fc:2 4 × 4 1 × 1 Batch size:10 224 × 224
Mishra et al. [55] CovAI-Net (Inception, DenseNet, Xception) Conv:8, pool:4, d:2, 7 × 7 3 × 3 Stride:2, Batch size:32 224 × 224
Shah et al. [108] CTNet-10 (DenseNet-169, VGG-16, ResNet-50, InceptionV3, VGG-19) Conv:5, pool:3, fc:3 Batch size:32 128 × 128 CTNet10, 224 × 224 VGG19
Sakib et al. [100] DL-CRC Batch size:8
Tang et al. [44] EDL-COVID 6 layers of CovidNet 3 × 3 1 × 1 Batch size:64
Saha et al. [94] EMCNet(AlexNet, VGG 16, Inception, ResNet-50) Conv:6, pool:5, ft:1, db:6, fc:2, 7 × 7 3 × 3 Batch size:32 224 × 224
Khan et al. [102] H3DNN(3D ResNets, C3D, 3DDenseNets, I3D, LRCN) Conv:9, pool: 6, incep:9,fc:2 7 × 7 3 × 3 Batch size:2 224 × 224
Gupta et al. [54] InstaCovNet-19 ( NasNet, (InceptionV3, Xception, ResNet101, MobileNetV2) 3 × 3 1 × 1 Batch size:16 224 × 224
Author Learning Rate Epoch Performance Activation function Optimizer Software
Panwar et al. [57] 47 Accuracy:95.61%, Sensitivity:94.04%, Specificity:95.86% Softmax Adam
Nath et al. [34] 0.001 15 Accuracy X-ray:99.68%, CT:71.81% ReLU SGDM, Adam, RmsProp Matlab 2019b
Kassani et al. [97] Accuracy:99% Keras package with Tensor flow
Hussain et al. [93] 0.0001 50 Accuracy: 99.1%, Precision:99.27%, Recall:98.17%, F1-score:98.51% Sigmoid, ReLU, Leaky ReLU Adam Google Colab Keras, Tensorflow 2.0
Gilanie et al. [135] 0.01 25 Accuracy:96.68%, Specificity:95.65%, Sensitivity:96.24% ReLU, Softmax MATLAB 2018b
Silva et al. [106] 0.001 20 Accuracy:87.6%, F1-score:86.19%, AUC:90.5% ReLu, Sigmoid Adam TensorFlow Keras
Turkoglu[49] 100 ReLUELM, PReLU-ELM, TanhRL
Horry et al. [101] 0.001, 0.00001 10 Precision: 86% for X-ray, 100% for LUS, 84% for CT scans Keras APIs with a TensorFlow 2
Dutta et al. [112] 0.00003 31 Accuracy:84% Sigmoid RMSprop Google Colab TensorFlow
Mertyüz et al. [51] 0.00001 20 Accuracy:96.90%, Sensitivity:95.45%, Specificity:100% ReLu, Sigmoid Anaconda Spyder, Keras
Ko et al. [38] 0.0001 optimal epoch Sensitivity:99.58%, Specificity:100.00%, Accuracy:99.87% SoftMax Adam TensorFlow package, Keras
Alazab et al. [67] 0.1 25 F-Measure:99% Pandas, Scikit, TensorFlow
Sharma et at [68] 0.0001 100 Accuracy:98%, Sensitivity:100%, Specificity:100%, Precision:100%, Relu Adam sklearn, tensorflow, keras
Apostolopoulos et al. [66] 10 Accuracy:96.78%, Sensitivity:98.66%, Specificity:96.46% Relu Adam
Wu et al. [115] 0.001 30 Accuracy:86.60%, Precision:96.20%, AUC:96.80% Softmax
Al-Waisy et al. [39] 0.01 10, 20 Accuracy:99.99%, Sensitivity:99.98%, Specificity:100%, Precision:100%, F1-score:99.99%, MSE:0.011%, RMSE:0.012% Adam, SGD
Haghanifar et al. [65] 0.0001 100 Accuracy:98.68%, F-score:94% Sigmoid Adam
Sarker et al. [69] 0.00001 Accuracy:94%, Precision:94%, Recall:94%, F-score:94% Softmax Adam
Wang et al. [70] 0.0002 22 Accuracy:93.3%, Sensitivity:100%, Precision:80% Adam Keras, Tensor- Flow
Mangal et al. [63] 0.0001 30 Accuracy:99.94%, Sensitivity100%, Precision:93.80% Sigmoid Adam
Javaheri et al. [40] 0.001 Accuracy:91.66%, Sensitivity:87.5%, Specificity:94%, AUC:95% Adam
Elkorany et al. [41] Recall:94.45%, Specificity:98.15%, Precision:94.42%, F1-score:94.4% Softmax
Tabik et al. [46] 0.0002 100 Accuracy:76.18%, Sensitivity:72.59%, Specificity:78.67%, F1-score:75.71%, Softmax SGD, Adam
Ucar et al. [103] 0.0004 35 Accuracy:98.26%, Specificity:99.13%, F1-score:98.25% ReLU MATLAB
Hemdan et al. [42] 0.001 50 Accuracy:90%, F1-score:91% SGD Keras package with TensorFlow2
Kedia et al. [48] 0.001 5 to 10 Accuracy:99.71%, ReLU, Softmax Adam Keras,TensorFlow2 Scikit-Learn
Ouchicha et al. [45] Adaptive LR 20 Accuracy:96.69%, Precision:96.72% Recall:96.84%, F1-score:96.68% Softmax Adam
Javor et al. [91] 17 Accuracy:95.6%, Sensitivity:99.3%, Specificity:75.8%
Ismael et al. [98] 0.01 10 Accuracy:94.7%, Sensitivity:91.00%, Specificity:98.89%, F1-score94.79%, AUC:99.90% ReLU SGDM MATLAB
Rohila et al. [124] 0.0001 100 Accuracy:94.9% ReLU Adam, SGD
Padma et al. [35] Accuracy: 99.2%, Sensitivity:99.1%, Specificity:98.8%, Precision:100% ReLU
Jain et al. [64] 0.001 14 Accuracy:97.97% LeakyReLU Adam
Anwar et al. [107] 0.0001 25 Accuracy:89.7, F1-score:89.6%, AUC:89.5% Adam Google colab
Sethi et al. [104] 200 Accuracy:98.6%, Specificity:99.3%, Precision:87.8%, Sensitivity:87.8%, F1-score:87.8% Adadelta Google Colab
Ying et al. [43] Accuracy:98.6%, Specificity:99.3%, Precision:87.8%, Sensitivity:87.8%
Jiang et al. [36] 0.0002 Accuracy:98.92%, Recall:97.80%, Precision:100.00%, F1-score:98.89% Adam TensorFlow
Yang et al. [92] 0.9 20 Accuracy:95%, Sensitivity:100%, Specificity:90%, F1-score:95%, AUC:99% SGD Sklearn 0.22.1
Serener et al. [111] Accuracy:89%, Sensitivity:98%, Specificity:86%, AUC:95% SGD
Basu et al. [28] 0.0001 100 Accurary:99% ReLU Adam, SGD TensorFlow, Keras
Wang et al. [95] 0.0001 50 Accuracy:96.1% Softmax
Arellano, Ramos [53] Sensitivity:89.47%, Specificity:100% ReLU Gradient Descent Pandas, seaborn, matplotlib, Keras,
Voulodimos et al. [96] Accuracy:99%, Recall:89%, Precision:91%, F1-Score:89% Keras, TensorFlow Google Colab
Chen et al. [1] Accuracy:96%, Sensitivity:98%, Specificity:94%, PPV:94.23%, NPV:97.92% Keras
Wu et al. [50] 0.00001 Accuracy:76%, AUC:81.9%, Sensitivity:81.1%, Specificity:61.5% rmsprop Keras 2.1.6
Minaee et al. [56] 0.0001 100 Sensitivity:98%, Specificity:92.9% SoftMax Adam
Demir[99] 0.001 150 Accuracy:100%, Sensitivity:100%, Specificity:100% SoftMax SGDM MATLAB 2020a
Perumal et al. [23] Accuracy:93.8% ReLu Adam
Sheykhivand et al. [52] 0.001 200 Accuracy:99.5%, Sensitivity100%, Specificity:99.02% Leaky-Relu RMSProp Keras, Tensorflow
Mishra et al. [55] 0.001 60 Accuracry:98.31%, Precision:100%, Sensitivity:96.74%, Specificity:100%, F1-score:98.34%, PPV:100% ReLU, Softmax Adam Tensorflow, Keras
Shah et al. [108] 0.0001 30 CTNet Accuracy:82.1%, VGG19 Accuracy: 94.52% Sigmoid RMSProp, Adam Google Colab
Sakib et al. [100] 0.001 100 Accuracy:93.94%, AUC:95.25% Leaky-Relu Adagrad Keras, TensorFlow
Tang et al. [44] 0:002 50 Accuracy:95%, Sensitivity:96.0%, PPV:94.1% TensorFlow 2.0.0
Saha et al. [94] 0.0001 50 Accuracy:98.91%, Precision:100%, Recall:97.82%, F1-score:98.89% ReLu
Khan et al. [102] 0.00001 1000 Accuracy:85% Softmax Adam Keras TensorFlow
Gupta et al. [54] Reduced L.R Accuracy:99.53%, Precision:100%, Recall:99%, F1-score:99% SoftMax, Relu Adam NumPy, Scikit-Learn, TensorFlow 2