Table 1.
Dataset of X-ray image from different sources
| Authors | Number of patients | X-ray image categories | Data sources | Results | Comments | |
|---|---|---|---|---|---|---|
| Narin et al. [75] | 8088 | Bacterial pneumonia | 2772 | GitHub repository | 99.7 | This GitHub repository consist of MERS, SARS and pneumonia X-ray or CT images |
| COVID-19 | 1023 | |||||
| Normal | 2800 | |||||
| Viral pneumonia | 1493 | |||||
| Apostolopoulos et al. [76] | 2870 | COVID-19 - | 1008 |
1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.GitHub repository 2.2.2.2.2.2.2.2.2.2.2.2.2.2.2.Radiological Society of North America (RSNA) 3.3.3.3.3.3.3.3.3.3.3.3.3.3.3.Radiopaedia 4.4.4.4.4.4.4.4.4.4.4.4.4.4.4.Italian Society of Medical 5.5.5.5.5.5.5.5.5.5.5.5.5.5.5.Interventional Radiology (SIRM) |
98.75 | Rescaled X-ray image to a size of 200 × 266 |
| Bacterial pneumonia | 1414 | |||||
| COVID-19 + | 448 | |||||
| Panwar et al. [77] | 5863 | Others | 5529 | Kaggle repository | 97% | Final input is provided as 224 × 224 × 3 image |
| COVID–19 + | 334 | |||||
| Jain et al. [78] | 6432 | Healthy person | 1583 | Kaggle repository | 97.97% | BBH hospital ward located at Rawalpindi |
| COVID-19 + | 576 | |||||
| Pneumonia infected | 4273 | |||||
| Majeed [3] | 6024 | Normal | 1575 | Radiological Society of North America (RSNA) [3] | 89.5% | Checking every image whether there is an artifact or not and the type of artifacts present in the images |
| Bacterial infection | 2771 | |||||
| Viral (non-COVID-19) | 1494 | |||||
| COVID19 | 184 | |||||
| S. Albahli and W. Albattah [79] | 2265 | COVID-19 | 850 | Kaggle and GitHub repository | 98.82 | Using X-rays, a small dataset was created to detect COVID-19 and non-COVID-19 infected people |
| non-COVID-19 | 500 | |||||
| pneumonia | 915 | |||||
| Heidari et al. [80] | 8474 | COVID–19 + | 415 | Georgetown University’s center [80] | 94.5% | X-ray images processing uses a threshold T = Vmin + 0.9 × (Vmax − Vmin) to segment the original image into a binary image |
| Infected pneumonia COVID-19 | 5179 | |||||
| Normal | 2880 | |||||
| Mohammadi et al. [81] | 545 | COVID-19 | 181 | GitHub repository | 90.0% | Different augmented parameter was performed with ranging value such as zooming, scaling, rotation |
| Normal | 364 | |||||
| Wang et al. [82] | 1102 | Normal | 565 | Kaggle and GitHub repository | 96.75% | Re-scaling, images normalization, and data augmentation are three types of image processing for X-rays |
| COVID-19 | 537 | |||||
| Fontanellaz [83] | 21,690 | Healthy | 10,311 |
1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.Italian Society of Medical 2.2.2.2.2.2.2.2.2.2.2.2.2.2.2.Interventional Radiology 3.3.3.3.3.3.3.3.3.3.3.3.3.3.3.Eurorad.org operated by the European Society of Radiology 4.4.4.4.4.4.4.4.4.4.4.4.4.4.4.Radiological Society of North America (RSNA) |
94.3% | It a robust X-ray-based diagnostic technique using deep learning |
| Pneumonia | 10,921 | |||||
| COVID-19 pneumonia | 458 | |||||