Table 1.

Comparison of models detecting COVID-19 cases, normal cases, and other chest diseases based on medical images (data derived from [22]).

Reference	Medical image	Disease detected, n			Accuracy (%)	Methodology		Gaps in classification
		COVID-19	Normal	Other chest diseases
Apostolopoulos and Mpesiana [14]	X-ray	224	504	700	93		Used transfer learning on VGG19. MobileNetV2, Inception, Xception, and InceptionResNetV2	Used only 3 classes: COVID-19, pneumonia, and other
Wang et al [23]	X-ray	53	8066	5526	92		Introduced COVID-Net—the first open-source COVID-19 detection system	Used only 3 classes: COVID-19, pneumonia, and normal
Narin et al [17]	X-ray	50	50	N/Aa	98		Used 5 pretrained networks and applied 3 binary classifications for 4 classes of chest x-rays	Used only 3 classes: normal, COVID-19, viral and bacterial pneumonia
Brunese et al [22]	X-ray	250	3520	2753	97		Defined 2 models based on VGG16: one to classify affected x-ray images from healthy ones and the other to classify COVID-19 from affected x-ray images. Then, they localized the affected areas.	Although they used x-ray images of most diseases, they used only 3 classes: COVID-19, healthy, and disease
Song et al [24]	CTb	777	708	N/A	86		Proposed DRE-Net and compared its performance with VGG-16, DenseNet, and ResNet	Used only 3 classes: COVID-19, bacterial pneumonia, and healthy
Zheng et al [25]	CT	313	229	N/A	90		Proposed DeCoVNet for classification	Used only 2 classes: COVID-19–positive and COVID-19–negative
Xu et al [26]	X-ray	219	175	224	86		Proposed ResNet-18 based CNNc network	Used only 3 classes: COVID-19, Influenza-A viral pneumonia, and normal
Ozturk et al [27]	X-ray	250	1000	500	92		Proposed DarkCovidNet	Used only 3 classes: COVID-19, pneumonia, and no findings
Ardakani et al [28]	CT	510	N/A	510	99		Used 10 CNN networks (ie, AlexNet and ResNet-101) for classification of 2 classes	Classified COVID-19 class from non–COVID-19 class
Li et al [18]	CT	1296	1325	1735	96		Proposed COV-Net for classifying 3 classes	Used only 3 classes: COVID-19, community-acquired pneumonia, and non-pneumonia
Abbas et al [15]	X-ray	105	80	11	95.12		Proposed DeTrac-ResNet18 CNN that uses Decompose, Transfer, and Compose architecture	Used only 3 classes: normal, COVID-19, and SARS
Chen et al [16]	CT	51	N/A	55	95.24		Used UNet++ along with Keras for segmentation and COVID-19 detection	Used only binary classification for COVID-19 detection

^aN/A: not applicable.

^bCT: computed tomography.

^cCNN: convolutional neural network.