Exploring Expression-Related Self-Supervised Learning and Spatial Reserve Pooling for Affective Behaviour Analysis

Abstract

Self-supervised learning (SSL) methods have gained attention for reducing dependence on labeled data. However, SSL methods are less investigated for facial expression recognition (FER), which requires expensive expression annotation, especially for large-scale video databases. In this paper, we explore an expression-related self-supervised learning (SSL) method called ContraWarping to perform expression classification in the 5th Affective Behavior Analysis in-the-wild (ABAW) competition. We also conduct a new spatial reserve pooling module to utilize all facial details for expression recognition. By evaluating on the Aff-Wild2 dataset, we demonstrate that ContraWarping outperforms existing supervised methods and other general SSL methods with only 0.7M trainable parameters and shows great application potential in the affective analysis area. Codes have been released at https://github.com/youqingxiaozhua/ABAW5.

Cite

Text

Xue et al. "Exploring Expression-Related Self-Supervised Learning and Spatial Reserve Pooling for Affective Behaviour Analysis." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2023. doi:10.1109/CVPRW59228.2023.00604

Markdown

[Xue et al. "Exploring Expression-Related Self-Supervised Learning and Spatial Reserve Pooling for Affective Behaviour Analysis." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2023.](https://mlanthology.org/cvprw/2023/xue2023cvprw-exploring/) doi:10.1109/CVPRW59228.2023.00604

BibTeX

@inproceedings{xue2023cvprw-exploring,
  title     = {{Exploring Expression-Related Self-Supervised Learning and Spatial Reserve Pooling for Affective Behaviour Analysis}},
  author    = {Xue, Fanglei and Sun, Yifan and Yang, Yi},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
  year      = {2023},
  pages     = {5701-5708},
  doi       = {10.1109/CVPRW59228.2023.00604},
  url       = {https://mlanthology.org/cvprw/2023/xue2023cvprw-exploring/}
}