Deep Fourier-based Exposure Correction with Spatial-Frequency Interaction (ECCV 2022)

Jie Huang+, Yajing Liu+, Feng Zhao*, Keyu Yan, Jinghao Zhang, Yukun Huang, Man Zhou, Zhiwei Xiong

*Equal Corresponding Authors

+Equal Contributions

University of Science and Technology of China (USTC)

Introduction

This repository is the official implementation of the paper, "Deep Fourier-based Exposure Correction with Spatial-Frequency Interaction", where more implementation details are presented.

0. Hyper-Parameters setting

Overall, most parameters can be set in options/train/train_Enhance.yml

1. Dataset Preparation

Create a .txt file to put the path of the dataset using

python create_txt.py

2. Training

python train.py --opt options/train/train_Enhance.yml

3. Inference

set is_training in "options/train/train_Enhance.yml" as False set the val:filelist as the validation set.

then

python train.py --opt options/train/train_Enhance.yml

Dataset

MSEC dataset (please refer to https://github.com/mahmoudnafifi/Exposure_Correction)

SICE dataset (I have uploaded it to https://share.weiyun.com/C2aJ1Cti)

Ours Results

MSEC dataset (https://drive.google.com/file/d/1AOuWujPffYaYJsDmkA8OcyU5S5a4zClP/view?usp=drive_link)

SICE dataset (https://drive.google.com/file/d/1_Ya8iyRqOGoOy10nAyTHZS-6dJcuLVhU/view?usp=drive_link)

Contact

If you have any problem with the released code, please do not hesitate to contact me by email (hj0117@mail.ustc.edu.cn).

Cite