In this repository we provide the official implementation of Self2Self with Dropout.

• Codename: Self2Self (CVPR 2020)
• Writers: Yuhui Quan ([email protected]); Mingqin Chen ([email protected]); Tongyao Pang ([email protected]); Hui Ji ([email protected])
• Institute: School of Computer Science and Engineering, South China University of Technology; Department of Mathematics, National University of Singapore

For more information please see:

• [paper]
• [supmat]
• [website]

Here is the list of libraries you need to install to execute the code:

• Python 3.6
• Tensorflow-gpu 1.14.0
• keras
• scikit-image
• scipy
• cv2 (opencv for python)

All of them can be installed via conda (anaconda), e.g.

conda install scikit-image

1. (Optional) Download the dataset you want and save them in './testsets/'.
2. Run the demo code in demo_denosing.py.

@InProceedings{Quan_2020_CVPR,
author = {Quan, Yuhui and Chen, Mingqin and Pang, Tongyao and Ji, Hui},
title = {Self2Self With Dropout: Learning Self-Supervised Denoising From Single Image},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2020}
}

For questions, please send an email to [email protected]