Skip to content
Globally and Locally Consistent Image Completion
Python
Branch: master
Clone or download
Latest commit 00dfe5b Aug 28, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
images update Feb 13, 2018
.DS_Store bug fix Mar 8, 2018
.gitignore remove unwanted files Aug 19, 2019
LICENSE update Feb 14, 2018
README.md Update README.md Aug 28, 2019
architecture.py update Aug 19, 2019
config.py update Aug 19, 2019
network.py update Aug 19, 2019
ops.py update Aug 19, 2019
test.py update Feb 13, 2018
train.py update Aug 19, 2019

README.md

Globally and Locally Consistent Image Completion

Tensorflow implementation of Globally and Locally Consistent Image Completion on celebA dataset.
Alt text

What's different from the paper

  • smaller image input size (128x128)
  • smaller patch sizes
  • less number of training iteration (500,000 iterations in the paper)
  • Adam optimizer used instead of Adadelta

Requirements

  • Opencv 2.4
  • Tensorflow 1.4

Folder Setting

-data
  -img_align_celeba
    -img1.jpg
    -img2.jpg
    -...

Train

$ python train.py 

To continue training

$ python train.py --continue_training=True

Test

$ python test.py --img_path=./data/test/test_img.jpg

Alt text

Use your mouse to erase pixels in the image.
When you're done, press ENTER.
Result will be shown in few seconds.

Results

Alt text

You can’t perform that action at this time.