Noisy-As-Clean

Install

To reproduce the results in the NAC paper:

 python nac_resnet_on_set12.py

The test is performed after the models are trained.

If you have pre-trained models, for AWGN denoising on Set12, you can modify the specific settings that need to test in test_nac_resnet_on_set12.py

 run test_nac_resnet_on_set12.py

The code for drawing the figures in ablation study is provided in the figures folder

 run plot_for_large_sigma.py

#Further comments: The code is mainly written by Yuan Huang.

Part of the code borrow from [Deep Image Prior]

Name		Name	Last commit message	Last commit date
Latest commit History 18 Commits
.idea		.idea
SomeISP_operator_python		SomeISP_operator_python
data		data
figures		figures
models		models
utils		utils
venv		venv
README.md		README.md
nac_dncnn_bsd68.py		nac_dncnn_bsd68.py
nac_dncnn_bsd68_blind.py		nac_dncnn_bsd68_blind.py
nac_dncnn_set12.py		nac_dncnn_set12.py
nac_dncnn_set12_blind.py		nac_dncnn_set12_blind.py
nac_mydncnn_bsd68.py		nac_mydncnn_bsd68.py
nac_mydncnn_bsd68_blind.py		nac_mydncnn_bsd68_blind.py
nac_mydncnn_set12.py		nac_mydncnn_set12.py
nac_mydncnn_set12_blind.py		nac_mydncnn_set12_blind.py
nac_resnet_bsd68.py		nac_resnet_bsd68.py
nac_resnet_bsd68_blind.py		nac_resnet_bsd68_blind.py
nac_resnet_set12.py		nac_resnet_set12.py
nac_resnet_set12_blind.py		nac_resnet_set12_blind.py
pipeline_denoising_withaug_post_square_input.py		pipeline_denoising_withaug_post_square_input.py
pipeline_s2s_cc_square_mydncnn.py		pipeline_s2s_cc_square_mydncnn.py
pipeline_s2s_cc_square_resnet.py		pipeline_s2s_cc_square_resnet.py
test_nac_resnet_on_set12.py		test_nac_resnet_on_set12.py