Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
You can not select more than 25 topics
Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
SuMuzi
b314d995fa
|
1 year ago | |
|---|---|---|
| .. | ||
| Datasets | 1 year ago | |
| pretrained_models | 1 year ago | |
| utils | 1 year ago | |
| MPRNet.py | 1 year ago | |
| README.md | 1 year ago | |
| config.py | 1 year ago | |
| data_RGB.py | 1 year ago | |
| dataset_RGB.py | 1 year ago | |
| evaluate_SIDD.m | 1 year ago | |
| generate_patches_SIDD.py | 1 year ago | |
| losses.py | 1 year ago | |
| test_DND.py | 1 year ago | |
| test_SIDD.py | 1 year ago | |
| train.py | 1 year ago | |
| training.yml | 1 year ago | |
README.md
Training
-
Download the Datasets
-
Generate image patches from full-resolution training images of SIDD dataset
python generate_patches_SIDD.py --ps 256 --num_patches 300 --num_cores 10
- Train the model with default arguments by running
python train.py
Evaluation
- Download the model and place it in
./pretrained_models/
Testing on SIDD dataset
- Download SIDD Validation Data and Ground Truth from here and place them in
./Datasets/SIDD/test/ - Run
python test_SIDD.py --save_images
Testing on DND dataset
- Download DND Benchmark Data from here and place it in
./Datasets/DND/test/ - Run
python test_DND.py --save_images
To reproduce PSNR/SSIM scores of the paper, run MATLAB script
evaluate_SIDD.m
