Table 1: Comparing AI and purely OSEM-based methods using conventional metrics and using AUC
| RMSE | SSIM | PSNR | AUC | |
| AI-based | 0.045 (0.045,0.046) | 0.869 (0.868,0.871) | 27.246 (27.134,27.353) | 0.735 (0.714,0.756) |
| Pure OSEM-based | 0.047 (0.047,0.048) | 0.851 (0.850,0.853) | 26.799 (26.689,26.899) | 0.744 (0.722,0.765) |
| Inference | AI-based has lower RMSE: Significantly Better (p < 0.001) | AI-based has higher SSIM: Significantly Better (p < 0.001) | AI-based has higher PSNR: Significantly Better (p < 0.001) | Both methods have similar AUC |