This is PyTorch implementation for Deep Guided Attention Network for Joint Denoising and Demosaicing in Real Image, Chinese Journal of Electronics 2024, by Tao Zhang, Ying Fu and Jun Zhang.
In this work, we propose a deep guided attention network for real image joint denoising and demosaicing, which considers the high signal-to-noise ratio and high sampling rate of green information for denoising and demosaicing, respectively.
- We present a deep guided attention network for real image JDD, that effectively considers the green channel characteristics of high SNR and high sampling rate in raw data.
- We propose a guided attention module to adaptively guide RGB image restoration by the information in green channel recovery branch.
- Training
python train_denoising.py
python train_JDD.py
- Testing
python test_JDD.py
If you find this work useful for your research, please cite:
@article{zhang2024DGAN,
title = {Deep Guided Attention Network for Joint Denoising and Demosaicing in Real Image},
author = {Zhang, Tao and Fu, Ying and Zhang, Jun},
journal = {Chinese Journal of Electronics},
volume = {33},
number = {E220414},
pages = {303},
year = {2024}
}