Skip to content
/ l2uwe Public

L^2UWE: A Framework for the Efficient Enhancement of Low-Light Underwater Images Using Local Contrast and Multi-Scale Fusion (CVPRW 2020)

Notifications You must be signed in to change notification settings

tunai/l2uwe

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

L^2UWE: Low-light underwater image enhancement

This repo contains the implementation of our work: L^2UWE: A Framework for the Efficient Enhancement of Low-Light Underwater Images Using Local Contrast and Multi-Scale Fusion, by Tunai Porto Marques and Alexandra Branzan Albu (presented at the 2020 CVPR Workshop NTIRE: New Trends in Image Restoration and Enhancement held in Seattle, June 15th).

L^2UWE uses a single image, local contrast information and a multi-scale fusion process to highlight visual features from the input that might have been originally hidden because of low-light settings. Presentation videos: 1 min. version 10 min. version

If L^2UWE proves to be useful to your work, we ask that you cite its related publications:

BibTeX

@inProceedings{Marques_2020_CVPR_Workshops,
title={L2UWE: A Framework for the Efficient Enhancement of Low-Light Underwater Images Using Local Contrast and Multi-Scale Fusion},
author={Porto Marques, Tunai and Branzan Albu, Alexandra},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
pages={538-539},
year={2020}}

@article{porto2019contrast,
title={A Contrast-Guided Approach for the Enhancement of Low-Lighting Underwater Images},
author={Porto Marques, Tunai and Branzan Albu, Alexandra and Hoeberechts, Maia},
journal={Journal of Imaging},
volume={5},
number={10},
pages={79},
year={2019},
publisher={Multidisciplinary Digital Publishing Institute} }

System requirements

  1. MATLAB
  2. Image Processing Toolbox

The framework was tested on MATLAB versions R2019b and R2020a.

Demo script

Open the "demo.m" script and point to your input image in the "imread" command. Some sample low-lighting underwater and aerial images are already provided in the "./data/" folder.

Once processed, the partial and final results are saved on the "./out/" folder.

Dataset

The dataset used in the development of L^2UWE, OceanDark, can be found in this repo at data/OceanDark2_0. Detailed information about it can be found at OceanDark.

Repo author

Tunai Porto Marques ([email protected]), website

About

L^2UWE: A Framework for the Efficient Enhancement of Low-Light Underwater Images Using Local Contrast and Multi-Scale Fusion (CVPRW 2020)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages