Model Level Ensemble for Facial Action Unit Recognition at the 3rd ABAW Challenge
In this paper, we present our latest work on Action Unit Detection, which is a part of the Affective Behavior Analysis in-the-wild (ABAW) 2022 Competition. Our proposed network is based on the IResnet100. First of all, We utilize feature pyramid networks (FPN) and single stage headless (SSH) to enlarge the receptive field and extract more facial texture features. Then we employ the ML-ROS data balancing and the BCE Loss plus Multi-label Loss to solve the multi-label imbalance problem. We also use three different models as the base model to fine-tune the Aff-Wild2 dataset. The pre-train backbones are the AU detection model, expression model and face recognition model. Finally, we adopt an ensemble methodology to get the final result. Our f1 score achieved 49.82 on the AU test set and ranked second in this challenge with a very small difference from the first team 49.89.