Table 2.
Artificial intelligence applied to thoracic Ct images for diagnosis.
| Author | Method | Subjects | Task classification | Results |
|---|---|---|---|---|
| Jin36 | U-Net++ | 723 covid 413 other |
Covid/other | S: 97.4% SP: 92.2% |
| Jin37 | CNN | 496covid 1385 other |
Covid/other | S: 94.1% SP: 95.5% |
| Li42 | COV-Net RESNET 50 |
468 covid 1551 nac 1303 other |
Covid/pneumonia/other | S: 90.0% SP: 96.0% AUC: 0.96 |
| Chen38 | U-Net++ | 50 covid 55 other |
Covid/other | S: 100% SP: 93.6% A: 95.2% |
| Wang41 | M. Inception | 79covid 180 viral pneumonia 15 covid |
Covid/pneumonia Prediction (pcr -) |
A: 82.5% S: 75% SP: 86% A: 85.2% |
| Xu44 | RESNet18 | 219 covid 224 influenza a 175normal |
Covid/influenza a/normal | A: 86.7% |
| Zheng39 | U-Net+3D DEEP NETWORK | 313covid 229 other |
Covid/other | S: 90.7% SP: 91.1% A: 90.8 AUC: 0.959 |
| Shi47 | V-Net RANDOM FOREST |
1658 covid 1027 pneumonia |
Covid/pneumonia | S: 90.7% SP: 83.3% AUC: 87.9% |
| Song43 | DRE-NET RESNET 5° |
88 covid 101 pneumonia 86 normal |
Covid/pneumonia/normal | A: 86% AUC: 0.95 |
| Bai40 | Efficient-Net B4 | 521 covid 665 pneumonia |
Covid/pneumonia | A: 96% AUC: 0.95 |