Action decouple multi-tasking for micro-expression recognition

Micro-expressions are brief, involuntary facial movements that reveal genuine emotions. However, extracting and learning features from micro-expressions poses challenges due to their short duration low intensity. To address this problem, we propose the ADMME (Action Decouple Multi-tasking for Micro-Expression Recognition) method. In our model, adopt a pseudo-siamese network architecture leverage contrastive obtain better representation of micro-expression motion features. During model training, utilize focal loss handle class imbalance issue in datasets. Additionally, introduce an AU Unit) detection task, which provides new inductive bias detection, enhancing model’s generalization robustness. Through five-class classification experiments conducted on CASMEII SAMM datasets, achieve accuracy rates 86.34% 81.28%, with F1 scores 0.8635 0.8168, respectively. These results validate effectiveness method recognition tasks. Furthermore, each component approach through series ablation experiments.