Facial Expression Recognition In-the-Wild with Deep Pre-trained Models.

Siyang Li,Yifan Xu,Huanyu Wu,Dongrui Wu,Yingjie Yin,Jiajiong Cao,Jingting Ding
DOI: https://doi.org/10.1007/978-3-031-25075-0_14
2022-01-01
Abstract:Facial expression recognition (FER) is challenging, when transiting from the laboratory to in-the-wild situations. In this paper, we present a general framework for the Learning from Synthetic Data Challenge in the 4th Affective Behavior Analysis In-The-Wild (ABAW4) competition, to learn as much knowledge as possible from synthetic faces with expressions. To cope with four problems in training robust deep FER models, including uncertain labels, class imbalance, mismatch between pretraining and downstream tasks, and incapability of a single model structure, our framework consists of four respective modules, which can be utilized for FER in-the-wild. Experimental results on the official validation set from the competition demonstrated that our proposed approach outperformed the baseline by a large margin.
What problem does this paper attempt to address?