Table 2.
Application of artificial intelligence in reducing image noise.
| Study | Year | Algorithm | Degree of image noise reduction | Image quality mean scores (AI group vs. contrast group) | Mean image noise (HU) (AI group vs. contrast group) | ERD mean (mm) (AI group vs. contrast group) | Degree of radiation dose reduction |
|---|---|---|---|---|---|---|---|
| Tatsugam et al. (27) | 2019 | DCNN | 20% | 3.58 vs. 2.96 | 18.5 vs. 23.0 | 16.7 vs.18.5 | 36% |
| Benz et al. (28) | 2020 | DCNN | 43% | 4.2–4.6 vs. 1.8–2.2 | 30 vs. 53 | NA | 65% |
| Hong et al. (29) | 2020 | CNN (U-net) | >20% | 3.65 vs. 2.45 | 52.64 vs. 67.22 | 0.9141 vs. 0.9589 | NA |
DCNN, deep convolutional neural network; CNN, convolutional neural network; ERD, edge rise distance; NA, not applicable.