Video Denoising on Tennis 240x352 150 frames (test)
37.81PSNR (sigma=5)ViDeNN-G
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ViDeNN-GModel version=Specifically trained for AWGN denoising2019.04 | 37.81 | 30.36 | 28.44 | |
| ViDeNNModel version=General model tackling multiple noise models2019.04 | 35.51 | 29.97 | 28 | |
| DnCNNMode=Blind model provided by authors2019.04 | 35.49 | 27.47 | 25.43 | |
| VBM4DComplexity=Low complexity setup, Mode=Blind version (noise estimation activated)2019.04 | 34.64 | 29.72 | 27.49 | |
| CBM3DMode=Blind evaluation (fixed sigma level 50)2019.04 | 27.04 | 26.37 | 25.62 |