The models provided in this section are meant for customizing your own denoiser! A summary of all the models will be available on Google Drive (we are working on it).
You can find the detailed step-by-step process in the demo.md.
We are currently dedicated to training an exceptionally capable network that can generalize well to various scenarios using only two data pairs! We will update this section once we achieve our goal. Stay tuned and look forward to it!
Or you can just use the following pretrained LED module for custumizing on your own cameras!.
| Method | Noise Model | Phase | Framework | Training Strategy | Additional Dgain (ratio) | Camera Model | Validation on | 🔗 Download Links | Config File |
|---|---|---|---|---|---|---|---|---|---|
| LED | ELD (5 Virtual Cameras) | Pretrain | UNet | PMN | 100-300 | - | - | [Google Drive] | [options/LED/pretrain/MM22_PMN_Setting.yaml] |
| LED | ELD (5 Virtual Cameras) | Pretrain | UNet | ELD | 100-300 | - | - | [Google Drive] | [options/LED/pretrain/CVPR20_ELD_Setting.yaml] |
| LED | ELD (5 Virtual Cameras) | Pretrain | UNet | ELD | 1-200 | - | - | [Google Drive] | [options/LED/pretrain/CVPR20_ELD_Setting_Ratio1-200.yaml] |
| LED | ELD (5 Virtual Cameras) | Pretrain | Restormer | ELD | 100-300 | - | - | [Google Drive] | [options/LED/other_arch/Restormer/LED+Restormer_Pretrain.yaml] |
| LED | ELD (5 Virtual Cameras) | Pretrain | NAFNet | ELD | 100-300 | - | - | [Google Drive] | [options/LED/other_arch/NAFNet/LED+NAFNet_Pretrain.yaml] |
The models provided in this section are used to replicate the metrics in our paper, and you can find a summary of all the models on Google Drive or Baidu Clould.
Notice that, all the models are trained for bayer format.
Training Strategydetermines the training strategy we use, you can find the specific method in their paper (ELD and PMN).
PMN* or ELD* for LED denotes the model is pretrained on that strategy.
| Method | Noise Model | Phase | Framework | Training Strategy | Additional Dgain (ratio) | Camera Model | Validation on | 🔗 Download Links | Config File |
|---|---|---|---|---|---|---|---|---|---|
| LED | ELD (5 Virtual Cameras) | Pretrain | UNet | PMN | 100-300 | - | - | [Google Drive] | [options/LED/pretrain/MM22_PMN_Setting.yaml] |
| LED | ELD (5 Virtual Cameras) | Pretrain | UNet | ELD | 100-300 | - | - | [Google Drive] | [options/LED/pretrain/CVPR20_ELD_Setting.yaml] |
| LED | ELD (5 Virtual Cameras) | Pretrain | UNet | ELD | 1-200 | - | - | [Google Drive] | [options/LED/pretrain/CVPR20_ELD_Setting_Ratio1-200.yaml] |
| LED | Real Noise (6 Pairs on SID SonyA7S2) | Finetune / Deploy | UNet | PMN* | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/LED/finetune/SID_SonyA7S2_MM22_PMN_Setting.yaml] |
| LED | Real Noise (6 Pairs on SID SonyA7S2) | Finetune / Deploy | UNet | ELD* | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/LED/finetune/SID_SonyA7S2_CVPR20_ELD_Setting.yaml] |
| LED | Real Noise (6 Pairs on SID SonyA7S2) | Finetune / Deploy | Restormer | ELD* | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/LED/other_arch/Restormer/LED+Restormer_Finetune.yaml] |
| LED | Real Noise (6 Pairs on SID SonyA7S2) | Finetune / Deploy | NAFNet | ELD* | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/LED/other_arch/NAFNet/LED+NAFNet_Finetune.yaml] |
| LED | Real Noise (24 Pairs on ELD SonyA7S2) | Finetune / Deploy | UNet | ELD* | 1-200 | SonyA7S2 | ELD SonyA7S2 | [Google Drive] | [options/LED/finetune/ELD_SonyA7S2_CVPR20_ELD_Setting.yaml] |
| LED | Real Noise (24 Pairs on ELD NikonD850) | Finetune / Deploy | UNet | ELD* | 1-200 | NikonD850 | ELD NikonD850 | [Google Drive] | [options/LED/finetune/ELD_NikonD850_CVPR20_ELD_Setting.yaml] |
| ELD | ELD (SonyA7S2) | Deploy | UNet | PMN | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/ELD/SID_SonyA7S2_MM22_PMN_Setting.yaml] |
| ELD | ELD (SonyA7S2) | Deploy | UNet | ELD | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/ELD/SID_SonyA7S2_CVPR20_ELD_Setting.yaml] |
| ELD | ELD (SonyA7S2) | Deploy | UNet | ELD | 1-200 | SonyA7S2 | ELD SonyA7S2 | [Google Drive] | [options/ELD/ELD_SonyA7S2_CVPR20_ELD_Setting.yaml] |
| ELD | ELD (NikonD850) | Deploy | UNet | ELD | 1-200 | NikonD850 | ELD NikonD850 | [Google Drive] | [options/ELD/ELD_NikonD850_CVPR20_ELD_Setting.yaml] |
| P-G | P-G (SonyA7S2) | Deploy | UNet | ELD | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/P-G/SID_SonyA7S2_CVPR20_ELD_Setting.yaml] |
| P-G | P-G (SonyA7S2) | Deploy | UNet | ELD | 1-200 | SonyA7S2 | ELD SonyA7S2 | [Google Drive] | [options/P-G/ELD_SonyA7S2_CVPR20_ELD_Setting.yaml] |
| P-G | P-G (NikonD850) | Deploy | UNet | ELD | 1-200 | NikonD850 | ELD NikonD850 | [Google Drive] | [options/P-G/ELD_NikonD850_CVPR20_ELD_Setting.yaml] |
| SID | Real Noise (SonyA7S2) | Deploy | UNet | ELD | 100-300 | SonyA7S2 | SID SonyA7S2 | [Google Drive] | [options/SID/SID_SonyA7S2_CVPR_ELD_Setting.yaml] |
Typecan be found inled/data/noise_utils/noise_generator.py.
p,g,t,r,q,cinNoise Typedenotes shot, read (gaussian), read (tukey-lambda), row, quant noise and black level error, respectively.
All the noise model can be found in Google Drive or Baidu Cloud.
| Type | Noise Model | Noise Type | Camera Model | 🔗 Download Links |
|---|---|---|---|---|
| VirtualNoisyPairGenerator | ELD (5 Virtual Cameras) | ptrqc | Virtual Camera | [Google Drive] |
| CalibratedNoisyPairGenerator | ELD (SonyA7S2) | ptrqc | SonyA7S2 | [Google Drive] |
| CalibratedNoisyPairGenerator | ELD (NikonD850) | ptrqc | NikonD850 | [Google Drive] |
| CalibratedNoisyPairGenerator | P-G (SonyA7S2) | pg | SonyA7S2 | [Google Drive] |
| CalibratedNoisyPairGenerator | P-G (NikonD850) | pg | NikonD850 | [Google Drive] |