Exploring Large-scale Unlabeled Faces to Enhance Facial Expression Recognition
Jun Yu, Zhongpeng Cai, Renda Li, Gongpeng Zhao, Guochen Xie, Jichao Zhu, Wangyuan Zhu, Qiang Ling, Lei Wang, Cong Wang, Luyu Qiu, Wei Zheng
Abstract
Facial Expression Recognition (FER) is an important task in computer vision and has wide applications in many fields. In this paper, we introduce our approach to the fifth Affective Behavior Analysis in-the-wild (ABAW) Competition which will be held in CVPR20223. For facial expression recognition task, there is an urgent need to solve the problem that the limited size of FER datasets limits the generalization ability of expression recognition models, resulting in ineffective performance. To address this problem, we propose a semi-supervised learning framework that utilizes unlabeled face data to train expression recognition models effectively. Our method uses a dynamic threshold module (DTM) that can adaptively adjust the confidence threshold to fully utilize the face recognition (FR) data to generate pseudo-labels, thus improving the model’s ability to model facial expressions. In the 5th ABAW Expression Classification Challenge, our method achieves good results on the Aff-Wild2 validation and test sets, demonstrating that large scale unlabeled faces can indeed improve the performance of face expression recognition.