Skip to content
This repository is an official PyTorch implementation of the paper "Progressive Feature Fusion Network for Realistic Image Dehazing". (ACCV 2018)
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
images release PFFNet Jun 15, 2019
model release PFFNet Jun 15, 2019
.DS_Store release PFFNet Jun 15, 2019
LICENSE release PFFNet Jun 15, 2019
README.md
data.py release PFFNet Jun 15, 2019
data_agument.py release PFFNet Jun 15, 2019
psnr_ssim.py release PFFNet Jun 15, 2019
test.py release PFFNet Jun 15, 2019
train.py release PFFNet Jun 15, 2019
utils.py release PFFNet Jun 15, 2019

README.md

PFFNet-PyTorch

This repository is an official PyTorch implementation of the paper "Progressive Feature Fusion Network for Realistic Image Dehazing".

Original link to this project: PFFNet.

Paper can be download from PFFNet

PFFNet is a solution for the NTIRE2018 image defogging challenge (20.549db for Indoor and 20.230db for Outdoor), final results could be found at NTIRE2018. We won the six place and awarded "Honorable Mention Award".

Improved version was accepted by ACCV2018.

All pretrained models can be found at: Here


Preparation

Using data_argument.py to enchance the datasets, it will produce below datasets

$ python dara_argument.py --fold_A=IndoorTrainHzay --fold_B=IndoorTrainGT --fold_AB=IndoorTrain 

IndoorTrain
\data   hazy image
\label  clear image

Train

Using default parameter to train

python train.py --cuda --gpus=4 --train=/path/to/train --test=/path/to/test --lr=0.0001 --step=1000

Test

python test.py --cuda --checkpoints=/path/to/checkpoint --test=/path/to/testimages

Results


Citation

If you use the code in this repository, please cite our paper:

@inproceedings{mei2018pffn,
title={Progressive Feature Fusion Network for Realistic Image Dehazing},
author={Mei, Kangfu and Jiang, Aiwen and Li, Juncheng and  Wang, Mingwen},
booktitle={Asian Conference on Computer Vision (ACCV)},
year={2018}
}
You can’t perform that action at this time.