Skip to content

Official repository of the Deep Image Debanding conference paper. This research project was done with Dr. Shahrukh Athar, Zhongling Wang, and Prof. Zhou Wang, and the work is published in ICIP 2022.

License

Notifications You must be signed in to change notification settings

RaymondLZhou/deepDeband

Repository files navigation

Deep Image Debanding

This repository contains the implementation for the following work:

Raymond Zhou, Shahrukh Athar, Zhongling Wang, and Zhou Wang, “Deep Image Debanding,” in IEEE International Conference on Image Processing (ICIP), Bordeaux, France, Oct. 2022, IEEE Xplore

Abstract

Banding or false contour is an annoying visual artifact whose impact negatively degrades the perceptual quality of visual content. Since users are increasingly expecting better visual quality from such content and banding leads to deteriorated quality-of-experience, the area of banding removal or debanding has taken paramount importance. Existing debanding approaches are mostly knowledge-driven, while data-driven debanding approaches remain surprisingly missing. In this work, we construct a large-scale dataset of 51,490 pairs of corresponding pristine and banded image patches, which enables us to make one of the first attempts at developing a deep learning based banding artifact removal method for images that we name deep debanding network (deepDeband). We also develop a bilateral weighting scheme that fuses patch-level debanding results to full-size images. Extensive performance evaluation shows that deepDeband is successful at greatly reducing banding artifacts in images, outperforming existing methods both quantitatively and visually.

Banding Example

Banded image

banded

Debanded (deepDeband-f)

debanded full

Debanded (deepDeband-w)

debanded weighted

Running the Model

  1. Install the prerequisites with pip install -r requirements.txt
  2. Place your input images in input/
  3. Navigate to the src directory with cd src
  4. Choose the version of deepDeband to run (f or w) with python deepDeband.py --version f or python deepDeband.py --version w
  5. The output images are located in output/deepDeband-f/ or output/deepDeband-w/

Dataset

The dataset of 51,490 pairs of matching banded and pristine image patches of size 256x256 is available here: DOI

Model

In case there are any issues with downloading the model in this repo through GitHub LFS, it is also available here: DOI. After downloading and extracting the zip file, place the deepDeband-f and deepDeband-w folders into pytorch-CycleGAN-and-pix2pix/checkpoints/.

About

Official repository of the Deep Image Debanding conference paper. This research project was done with Dr. Shahrukh Athar, Zhongling Wang, and Prof. Zhou Wang, and the work is published in ICIP 2022.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages