CT Denoising on GBA-LDCT (test)
43.38PSNRRelativeFlow
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| RelativeFlowType=SimSGL, Category=MID, Backbone=2D Guided U-Net, Inference Steps=32026.04 | 43.38 | 93.59 | 0.7 | 4.38 | |
| RED-CNNType=SimSDL, Category=MID, Backbone=Specific architectural contribution2026.04 | 43.12 | 92.15 | 0.74 | 5.18 | |
| Flow MatchingType=SimSGL, Category=NID, Backbone=2D Guided U-Net, Inference Steps=32026.04 | 42.05 | 92.88 | 0.83 | 4.51 | |
| SwinIRType=SimSDL, Category=NID, Backbone=2D U-Net2026.04 | 41.78 | 93.12 | 1.13 | 4.62 | |
| Noise2SelfType=SSL, Category=NID, Backbone=2D U-Net2026.04 | 36.45 | 81.55 | 1.52 | 19.85 | |
| Noise2SimType=SSL, Category=MID, Backbone=2D U-Net2026.04 | 36.35 | 88.71 | 1.52 | 9.18 | |
| DDIMType=SimSGL, Category=NID, Backbone=2D Guided U-Net, Inference Steps=32026.04 | 35.71 | 89.58 | 1.64 | 11.21 | |
| IPDMType=SimSGL, Category=MID, Backbone=2D Guided U-Net, Inference Steps=32026.04 | 31.31 | 92.42 | 2.82 | 5.97 |