Low-dose CT Denoising on Mayo 2016 (test) using Perceptual Metrics
33.21MANIQA ScoreRED-CNN
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| RED-CNNResolution=512x5122026.06 | 33.21 | 57.71 | 58.53 | |
| DeVARResolution=512x512, Base Model=GPT-2 style transformer, Base Model Blocks=16, Residual Refiner Blocks=62026.06 | 33.12 | 62.58 | 58.68 | |
| CTformerResolution=512x5122026.06 | 32.82 | 57.55 | 56.48 | |
| ASCONResolution=512x5122026.06 | 31.96 | 55.06 | 55 | |
| DU-GANResolution=512x5122026.06 | 31.78 | 58.09 | 59.5 | |
| CoreDiffResolution=512x5122026.06 | 30.49 | 54.76 | 56.59 |