Contrastive Learning for Compact Single Image Dehazing, CVPR2021. Official Pytorch based implementation.
- model
- CR loss
- pretrained models
https://github.com/Booooooooooo/AECRNet-MindSpore by @wyb
Pretrained models:
https://pan.baidu.com/s/13crsXwwhkI5A3MlHtPihuA password: xhyi
.
├── create_dataset.py
├── datasets
│ ├── ITS_test
│ └── ITS_train
├── data_utils
│ ├── DH.py
│ ├── ITS_h5.py
│ ├── NH.py
│ └── __pycache__
├── img
│ ├── aecrnet.png
│ ├── example.png
│ ├── performance.png
│ └── trade_off.png
├── ITS_v2
│ ├── clear
│ ├── clear.zip
│ ├── hazy
│ ├── hazy.zip
│ ├── trans
│ └── trans.zip
├── logs
│ ├── ITS_train_cdnet_test
│ ├── its_train_ffa_test
│ └── ITS_train_ffa_test
├── logs_train
│ ├── args_ITS_train_cdnet_test.txt
│ ├── args_its_train_ffa_test.txt
│ ├── args_ITS_train_ffa_test.txt
│ └── ITS_train_cdnet_test.txt
├── metrics.py
├── models
│ ├── AECRNet.py
│ ├── CR.py
│ ├── DCNv2
│ ├── DCNv2.zip
│ ├── deconv.py
│ └── __pycache__
├── numpy_files
├── option.py
├── README.md
├── requirements
├── samples
│ ├── ITS_train_cdnet_test
│ ├── its_train_ffa_test
│ └── ITS_train_ffa_test
├── test.py
├── train_aecrnet.py
└── trained_models
├── ITS_train_cdnet_test.pk
└── ITS_train_cdnet_test.pk.best
refer to: https://sites.google.com/view/reside-dehaze-datasets/reside-v0
ITS (Indoor Training Set):
(Dropbox): http://t.cn/RHjBQIV
(Baidu Yun):https://pan.baidu.com/s/16rm4zUF8uVRs3Ux5T9CMMA Passward: tqyh
put on ./ITS_v2
, and run python create_dataset.py
to create h5 files.