Attention Based Relation Network for Facial Action Units Recognition

TL;DR

This paper introduces ABRNet, an attention-based model that automatically learns and refines facial action unit relations for improved recognition accuracy, surpassing previous methods on benchmark datasets.

Contribution

The paper presents a novel attention-based relation network that automatically captures and refines AU relations without predefined rules, enhancing facial AU recognition.

Findings

Achieves state-of-the-art performance on DISFA and DISFA+ datasets.

Effectively models AU relations with relation dropout and loss strategies.

Improves robustness of AU features through self-attention fusion.

Abstract

Facial action unit (AU) recognition is essential to facial expression analysis. Since there are highly positive or negative correlations between AUs, some existing AU recognition works have focused on modeling AU relations. However, previous relationship-based approaches typically embed predefined rules into their models and ignore the impact of various AU relations in different crowds. In this paper, we propose a novel Attention Based Relation Network (ABRNet) for AU recognition, which can automatically capture AU relations without unnecessary or even disturbing predefined rules. ABRNet uses several relation learning layers to automatically capture different AU relations. The learned AU relation features are then fed into a self-attention fusion module, which aims to refine individual AU features with attention weights to enhance the feature robustness. Furthermore, we propose an AU relation dropout strategy and AU relation loss (AUR-Loss) to better model AU relations, which can further improve AU recognition. Extensive experiments show that our approach achieves state-of-the-art performance on the DISFA and DISFA+ datasets.