SR artifact segmentation on DeSRA
19.07F1 ScoreMolodetskikh et al.
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Molodetskikh et al.Threshold (t)=0.3, Requires Reference=Yes2026.05 | 19.07 | 48.66 | |
| DeSRAThreshold (t)=0.3, Requires Reference=Yes2026.05 | 17.52 | 52.77 | |
| DISTSThreshold (t)=0.25, Requires Reference=Yes2026.05 | 16.37 | 49.19 | |
| LDLThreshold (t)=0.005, Requires Reference=Yes2026.05 | 16.18 | 37.24 | |
| Mask2FormerTraining Strategy=From Scratch, Training Data=Q-Ground + Open Images [23] + SR-Ground, Requires Reference=No2026.05 | 15.38 | 37.37 | |
| Mask2FormerTraining Strategy=Fine-tuned Baseline, Training Data=Q-Ground + Open Images [23] + SR-Ground, Requires Reference=No2026.05 | 15.37 | 37.6 | |
| Mask2FormerTraining Strategy=From Scratch, Training Data=Q-Ground + Open Images [23], Requires Reference=No2026.05 | 14.63 | 34.37 | |
| TOPIQThreshold (t)=0.5, Requires Reference=No2026.05 | 1.6 | 4.24 | |
| PaQ-2-PiQThreshold (t)=65, Requires Reference=No2026.05 | 1.56 | 3.05 |