Table 1 Comparison of the open-source dataset sources of COVID-19 imaging. (Diag means diagnosis and Seg means segmentation)
From: A deep adversarial model for segmentation-assisted COVID-19 diagnosis using CT images
Dataset | Data type | Size (Cov/NonCov) | Task-Diag | Task-Seg |
|---|---|---|---|---|
COVID-19 X-ray collection [44] | X-rays | 229 | \(\checkmark\) | – |
COVID-19 CT collection [44] | CT volume | 20 | \(\checkmark\) | – |
COVID-CT-dataset [45] | CT image | 349/1000 | \(\checkmark\) | – |
COVID-19 patient lungs [46] | X-rays | 70/28 | \(\checkmark\) | – |
COVID-19 radiography [47] | X-rays | 1143/1314/1345 (viral pneumonia) | \(\checkmark\) | – |
SARS-CoV-2 CT-scan [48] | CT image | 1252/1230 | \(\checkmark\) | – |
COVID-19 CT segmentation [43] | CT image | 110/0 | – | \(\checkmark\) |
COVID-SemiSeg [12] | CT image | 1700/0 | – | \(\checkmark\) |
COVID-19 CT lung and infection segmentation [49] | CT image | 20/0 | – | \(\checkmark\) |
COVID-19 radiologist dataset | CT image | 100/93 | \(\checkmark\) | \(\checkmark\) |