Artifact Restoration on Histology Images (test)
0L2 ErrorArtiFusion (Full Settings)
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ArtiFusion (Full Settings)Denoising network=Swin-Transformer, Time token scheme=concatenation2023.07 | 0 | 0.2465 | 0.986 | 48.08 | 0.8216 | 55.43 | |
| ArtiFusion (Add)Time token scheme=direct summation2023.07 | 0.0001 | 0.2499 | 0.985 | 47.79 | 0.8184 | 54.76 | |
| ArtiFusion (U-Net)Denoising network=conventional U-Net2023.07 | 0.0001 | 0.2508 | 0.985 | 47.61 | 0.8173 | 54.59 | |
| CycleGAN (#1)Training samples=entire dataset2023.07 | 0.0001 | 0.5583 | 0.9656 | 42.37 | 0.7188 | 51.42 | |
| CycleGAN (#2)Training samples=half dataset2023.07 | 0.0002 | 0.5936 | 0.9622 | 42.12 | 0.7162 | 50.21 |