Learning Consistent Global-Local Representation for Cross-Domain Facial Expression Recognition

Implementation of papers:

• Learning Consistent Global-Local Representation for Cross-Domain Facial Expression Recognition
  2022 26th International Conference on Pattern Recognition (ICPR)
  Yuhao Xie, Yuefang Gao, Jiantao Lin, Tianshui Chen

Environment

Ubuntu 22.04.2 LTS, python 3.8.10, PyTorch 1.9.0

Datasets

Application website： SFEW 2.0, FER2013, ExpW, RAF.

Trained Model ( Table Ⅰof paper )

Backbone \ Target Datasets	FER2013	ExpW	SFEW
ResNet-50	1638244076	1638115309	1638294758
MobileNet-v2	1648861856	1648861886	1648878815

The code for each configuration can be found in this link using the file timestamp in the table above .

Usage

cd code
bash TrainOnSourceDomain.sh     # First step
bash TransferToTargetDomain.sh  # Second step

Citation

@INPROCEEDINGS{9956069,
  author={Xie, Yuhao and Gao, Yuefang and Lin, Jiantao and Chen, Tianshui},
  booktitle={2022 26th International Conference on Pattern Recognition (ICPR)}, 
  title={Learning Consistent Global-Local Representation for Cross-Domain Facial Expression Recognition}, 
  year={2022},
  volume={},
  number={},
  pages={2489-2495},
  doi={10.1109/ICPR56361.2022.9956069}}

Contributors

For any questions, feel free to open an issue or contact us:

• [email protected]