Table 2.
Multisource dataset used after preprocessing.
| Source | ChestX-ray8[2] |
Kaggle Pneumonia[3] | Covid19[4] | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Labels | Criteria | Atelectasis | Cardiomegaly | Effusion | Infiltration | Healthy | Pneumothorax | Viral | Bacterial | Covid-19 |
| Sample Size | Actual | 4215 | 1093 | 3955 | 9547 | 60,361 | 2194 | 2780 | 1493 | 6054 |
| Eliminated | 774 | 351 | 557 | 1346 | 4731 | 832 | 586 | 434 | 1893 | |
| Remaining | 3441 | 742 | 3398 | 8201 | 55,630 | 1362 | 2194 | 1059 | 4161 | |
| Balanced class | Real | 1100 | 742 | 1100 | 1100 | 1100 | 1100 | 1100 | 1059 | 1100 |
| Augmented | 0 | 358 | 0 | 0 | 0 | 0 | 0 | 41 | 0 | |