Low-light Image Enhancement

ArXiv

Official Matlab implementation of the paper: "Low-light Image Enhancement using Gaussian Process for Features Retrieval"

Signal Processing: Image Communication 2019

Released on June 1, 2019

Description

In this work, we propose to model low-light enhancement as a set of localized functions using Gaussian Process that is trained at runtime using data from a simple Convolutional Neural Network (CNN) to provide the necessary feature information as reference. The CNN is in turn trained using large amount of synthetic data, based upon the luminance distribution of real world low-light images to learn the relationship between features and pixels.

Citation

Please cite the following paper if you use this repository in your reseach:

@article{loh2019low,
  title={Low-light image enhancement using Gaussian Process for features retrieval},
  author={Loh, Yuen Peng and Liang, Xuefeng and Chan, Chee Seng},
  journal={Signal Processing: Image Communication},
  volume={74},
  pages={175--190},
  year={2019},
  publisher={Elsevier}
}

Dependencies

The codes are implemented in MATLAB using the prior version of MatConvNet and the Gaussian Process for Machine Learning toolboxes.

Installation and Running

1. Extract GPR_v1.1.zip

2. Extract matconvnet-1.0-beta20.tar.gz

   • run vl_compilenn
   • run vl_setupnn

   (for problems installing the toolbox, please refer to MatConvNet)

3. Run demo.m

Feedback

Suggestions and opinions of this work (both positive and negative) are greatly welcomed. Please contact the authors by sending an email to lexloh2009 at hotmail.comor cs.chan at um.edu.my.

License

The project is open sourced under BSD-3 license (see the LICENSE file). Codes can be used freely only for academic purposes.

For commercial purpose usage, please contact Dr. Chee Seng Chan at cs.chan at um.edu.my