Crack Restoration on Digitized paintings (test)
64.87SSIMLearning-based Segmentation
Evaluation Results
| Method | Links | |
|---|---|---|
| Learning-based SegmentationDetection method=Learning-based, Filling method=AD2026.02 | 64.87 | |
| Eq. (3)Guided Logit=true, LoRA=true, Mask Loss=true, Filling method=AD2026.02 | 64.87 | |
| Learning-based SegmentationDetection method=Learning-based, Filling method=MTM2026.02 | 61.24 | |
| Grassfire AlgorithmDetection method=Grassfire, Filling method=AD2026.02 | 56.03 | |
| Pik-FixTask=Specialized inpainting2026.02 | 55.71 | |
| Grassfire AlgorithmDetection method=Grassfire, Filling method=MTM2026.02 | 53.18 | |
| Wan et al.Task=End-to-end restoration2026.02 | 51.35 | |
| Eq. (4)Guided Logit=true, Filling method=AD2026.02 | 49.4 | |
| Eq. (1)Guided Logit=true, LoRA=true, Filling method=AD2026.02 | 48.67 | |
| SegFormerFilling method=AD2026.02 | 43.4 |