"Globally and Locally Consistent Image Completion" with Keras
Clone or download
Latest commit 5f0ecc4 Apr 12, 2018
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data default directory Jan 24, 2018
output update result image Apr 11, 2018
.gitignore default directory Jan 24, 2018
README.md add data structure Feb 3, 2018
model.py modify model function Mar 23, 2018
requirements.txt modify tensorflow keras version Mar 25, 2018
train.py modify model function Mar 23, 2018

README.md

Globally and Locally Consistent Image Completion

This is an Keras implementation of "Globally and Locally Consistent Image Completion".

Install python libraries

pip install -r requirements.txt

Prepare dataset

Please download appropriate image dataset and put it under the "data" folder. For example, please put the images as follows.

data
└── place365
    ├── Places365_00000001.jpg
    ├── Places365_00000002.jpg
    ├── Places365_00000003.jpg
    ├── Places365_00000004.jpg
    ├── Places365_00000005.jpg
    ├── Places365_00000006.jpg
    ├── Places365_00000007.jpg
    ├── Places365_00000008.jpg
    ...

Train

python train.py

Result

result