README.md

DeepFillv2 Test Images Guide

Test stages of Gated Convolution (DeepFillv2) inpainting method are explained.

Requirements

• Install python3.
• Install tensorflow (tested on Release 1.3.0, 1.4.0, 1.5.0, 1.6.0, 1.7.0).

I used Python 3.7, Tensorflow 1.15, CUDA 10

Test Stages

• Install software requirements
• Clone the deepfillv2 github project on your computer = git clone https://github.com/JiahuiYu/generative_inpainting
• Install tensorflow toolkit neuralgym (run pip install git+https://github.com/JiahuiYu/neuralgym)
• Generate masked images (link). I used NVIDIA Irregular Mask Dataset: Testing Set.
• Generate test commands file (link). (The system takes 2 inputs to test : masked image and mask image. Pay attention the masked image and mask image paths while generating test command file.)
• Run the commands file in your computer.

Examples

From celeba_hq dataset. First column shows masked image, second column shows mask image and third column shows inpainting result.

Citing

@article{yu2018generative,
  title={Generative Image Inpainting with Contextual Attention},
  author={Yu, Jiahui and Lin, Zhe and Yang, Jimei and Shen, Xiaohui and Lu, Xin and Huang, Thomas S},
  journal={arXiv preprint arXiv:1801.07892},
  year={2018}
}

@article{yu2018free,
  title={Free-Form Image Inpainting with Gated Convolution},
  author={Yu, Jiahui and Lin, Zhe and Yang, Jimei and Shen, Xiaohui and Lu, Xin and Huang, Thomas S},
  journal={arXiv preprint arXiv:1806.03589},
  year={2018}
}