SymmFCNet
SymmFCNet copied to clipboard
Learning Symmetry Consistent Deep CNNs for Face Completion
SymmFCNet
Torch implementation for Learning Symmetry Consistent Deep CNNs for Face Completion
<B>(New)
</B>Pytorch Version can be found here.
+ Please note that there are some different implementations between torch and pytorch version.
+ Please refer to pytorch version (https://github.com/csxmli2016/SymmFCNet_pytorch).
SymmFCNet framework (Torch Version)
Overview of our SymmFCNet. <B>Red</B>, <B>green</B> and <B>blue</B> lines represent the pixel-wise correspondence between the input and the flip image.
- <B>Red</B>: missing pixels (input) to non-occluded pixels (flip);
- <B>Green</B>: missing pixels (input) to missing pixels (flip);
- <B>Blue</B>: remaining pixels (input) to remaining pixels (flip).
![](https://github.com/csxmli2016/SymmFCNet/raw/master/Imgs/Pipeline/SymmFCNet.png)
Models
Download the pre-trained model with the following url and put it into ./checkpoints/.
Testing
th test.lua
Completion results
Type | Input | Results | Ground-Truth |
---|---|---|---|
Regular Mask |
![]() |
![]() |
![]() |
Irregular Mask |
![]() |
![]() |
![]() |
Real Occlusion |
![]() |
![]() |
![]() |
Requirements and Dependencies
Citation
@article{li2018learning,
title={Learning Symmetry Consistent Deep CNNs for Face Completion},
author={Li, Xiaoming and Liu, Ming and Zhu, Jieru and Zuo, Wangmeng and Wang, Meng and Hu, Guosheng and Zhang, Lei},
journal={arXiv preprint arXiv:1812.07741},
year={2018}
}