6zhc / DECUN

DECUN -- Deep Convergent Unrolling for Non-blind deblurring

DECUN -- Deep Convergent Unrolling for Non-blind deblurring

This repo contains the implementation of DECUN and some useful files to reuse its building blocks.

Structure of the code

1. data contains data samples for testing. The full data can be download from Here.

2. trained_models contains trained for testing.

3. test.py is the main test file to be run.

4. option.py contains the running option for test.py .

5. decovNet.py contains the DECUN network .

Setting up the environment

All required packages are found in requirements.txt.

• Creating the conda env for "DECUN"

conda create env -n "DECUN" python=3.9

• Activate "DECUN" conda env

conda activate DECUN

• Install PyTorch

pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu116

• Install required library

pip install -r requirements.txt

Running the Test file.

For more running option can see in the option.py

python test.py

If you want to use this code, please cite our paper as

@ARTICLE{Yanan24, author = {Yanan Zhao and Yuelong Li and Haichuan Zhang and Vishal Monga and Yonina C. Eldar}, title = {Deep, convergent, unrolled half-quadratic splitting for image deconvolution}, journal={arXiv preprint arXiv:2402.12872}, year = {2024}, }