fayolle / bbDeblur_py

Implementation (in Python) of iterative methods for black-box image deblurring

Reverse filtering for deblurring

Python code implementing five reverse filtering (or defiltering) schemes for noisy image deblurring described in the following paper.

Dependencies

The code depends on a few common Python libraries (NumPy, ...). Install them via pip and the provided requirements.txt file. For example

$ python -m venv bbdeblur_venv
$ ./bbdeblur_venv/Scripts/activate
$ pip install -r requirements.txt

How to run?

One possibility is to execute the scripts in the sub-directory src

$ cd src
$ python test_color.py

Another possibility is to use the Python notebook, provided in the sub-directory notebooks. The notebook works on Google Colab (with the default installation).

Notes

See also the original (and more complete) MATLAB implementation in this repo.

Reference

Link to the paper where the methods were introduced. The corresponding bibtex entry is

@article{Belyaev2022,
title = {Black-box image deblurring and defiltering},
author = {Belyaev, Alexander G. and Fayolle, Pierre-Alain},
journal = {Signal Processing: Image Communication},
pages = {116833},
year = {2022},
issn = {0923-5965},
doi = {https://doi.org/10.1016/j.image.2022.116833},
url = {https://www.sciencedirect.com/science/article/pii/S0923596522001242},
keywords = {Deblurring, Defiltering, Reverse filtering},
}