UAU-Net: Uncertainty-aware Representation Learning and Evidential Classification for Facial Action Unit Detection

2026-04-22

AI Summary

This study addresses the challenge of heterogeneous uncertainties in facial action unit (AU) detection, arising from visual noise, individual variability, and ambiguous inter-AU relationships, which—compounded by severe label imbalance—often lead classifiers to overconfident predictions. To tackle this, the authors propose UAU-Net, a novel framework that jointly models representation- and decision-level uncertainties for the first time. Specifically, it introduces a conditional variational autoencoder-driven multi-scale probabilistic feature extraction module (CV-AFE) at the representation stage and an asymmetric Beta evidential neural network (AB-ENN) tailored for highly imbalanced binary labels at the decision stage. Experimental results on the BP4D and DISFA datasets demonstrate that the proposed method achieves state-of-the-art performance, significantly enhancing model robustness and prediction reliability.

Technology Category

• Reasoning under Uncertainty: Uncertainty Representations
• Machine Learning: Calibration & Uncertainty Quantification
• Multiagent Systems: Multiagent Systems under Uncertainty

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Facial action unit (AU) detection remains challenging because it involves heterogeneous, AU-specific uncertainties arising at both the representation and decision stages. Recent methods have improved discriminative feature learning, but they often treat the AU representations as deterministic, overlooking uncertainty caused by visual noise, subject-dependent appearance variations, and ambiguous inter-AU relationships, all of which can substantially degrade robustness. Meanwhile, conventional point-estimation classifiers often provide poorly calibrated confidence, producing overconfident predictions, especially under the severe label imbalance typical of AU datasets. We propose UAU-Net, an Uncertainty-aware AU detection framework that explicitly models uncertainty at both stages. At the representation stage, we introduce CV-AFE, a conditional VAE (CVAE)-based AU feature extraction module that learns probabilistic AU representations by jointly estimating feature means and variances across multiple spatio-temporal scales; conditioning on AU labels further enables CV-AFE to capture uncertainty associated with inter-AU dependencies. At the decision stage, we design AB-ENN, an Asymmetric Beta Evidential Neural Network for multi-label AU detection, which parameterizes predictive uncertainty with Beta distributions and mitigates overconfidence via an asymmetric loss tailored to highly imbalanced binary labels. Extensive experiments on BP4D and DISFA show that UAU-Net achieves strong AU detection performance, and further analyses indicate that modeling uncertainty in both representation learning and evidential prediction improves robustness and reliability.

Problem

• Facial Action Unit Detection
• Uncertainty
• Label Imbalance
• Robustness
• Confidence Calibration

Innovation

• Uncertainty-aware representation learning
• Evidential classification
• Conditional VAE
• Asymmetric Beta distribution
• Facial action unit detection