Generative Image Inpainting with
Contextual Attention and Gated Convolution




DeepFill Demo: jhyu.me/deepfill

Notes:

  1. Results are direct outputs from trained generative neural networks. No post-processing steps are applied.
  2. Model is trained on CelebA-HQ and Places2 (with randomly sampling 2k as validation set for demo).
  3. Demo is for research purposes only.

Some examples are shown here. Tag #deepfill.

Now I am watching you Smile😁 Remove watermark. Edit bangs Swap eyes No mustache


Brush Width:

15

YouTube Video Demo

Best viewed with highest resolution 1080p.



Citation

@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}
}