Spatial-temporal Transformer for Affective Behavior Analysis
This work addresses affective computing for real-world applications, but appears incremental as it builds on existing Transformer methods for a specific competition.
The paper tackled affective behavior analysis in-the-wild by proposing a Transformer Encoder with Multi-Head Attention framework to learn spatial and temporal features, achieving results that demonstrate effectiveness on the Aff-Wild2 dataset.
The in-the-wild affective behavior analysis has been an important study. In this paper, we submit our solutions for the 5th Workshop and Competition on Affective Behavior Analysis in-the-wild (ABAW), which includes V-A Estimation, Facial Expression Classification and AU Detection Sub-challenges. We propose a Transformer Encoder with Multi-Head Attention framework to learn the distribution of both the spatial and temporal features. Besides, there are virious effective data augmentation strategies employed to alleviate the problems of sample imbalance during model training. The results fully demonstrate the effectiveness of our proposed model based on the Aff-Wild2 dataset.