Skip to content

fatalfeel/DeblurGAN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

98 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeblurGAN

arXiv Paper Version

Pytorch implementation of the paper DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networks.

Our network takes blurry image as an input and procude the corresponding sharp estimate, as in the example:

The model we use is Conditional Wasserstein GAN with Gradient Penalty + Perceptual loss based on VGG-19 activations. Such architecture also gives good results on other image-to-image translation problems (super resolution, colorization, inpainting, dehazing etc.)

Prerequisites data

  • cd ~/DeblurGAN
  • ./install_data.sh
  • or
  • bash -x ./install_data.sh

Train

  • step 1 open terminal
  • step 2 pip3 install visdom
  • step 3 python3 -m visdom.server
  • step 4 open another terminal
  • step 5 cd ~/DeblurGAN
  • step 6 python3 ./train.py --dataroot ./data/combined --resize_or_crop crop --cuda True
  • If you do not want to use visdom.server then skip step 1~6 and use these commands
  • python3 ./train.py --dataroot ./data/combined --resize_or_crop crop --display_id -1 --cuda True
  • [----------Resume training--------------]
  • python3 ./train.py --dataroot ./data/combined --resize_or_crop crop --display_id -1 --cuda True --resume True
  • [----------FPN101 and Wgan-gp------]
  • python3 ./train.py --dataroot ./data/combined --resize_or_crop crop --display_id -1 --cuda True --which_model_netG FPN101 --gan_type wgan-gp

Test

  • python3 ./test.py --dataroot ./data/blurred --model test --dataset_mode single --cuda True
  • [----------FPN101----------]
  • python3 ./test.py --dataroot ./data/blurred --model test --dataset_mode single --cuda True --which_model_netG FPN101

Model trained 2000 times

https://drive.google.com/file/d/1vGiqFXa177sCGHEuKhDKQ0VxvZZ2qpZg

Help you understand code

http://fatalfeel.blogspot.com/2013/12/deblurgan-image-synthesis-and-analysis.html

About

DeblurGAN simplize easy to lean

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published