AffectFlow-DINO: Uncertainty-Aware Multi-Task Affect Estimation via Conditional Rectified Flow
For researchers in affective computing and facial behavior analysis, this work provides a novel generative approach to handle ambiguity and class imbalance in in-the-wild affect estimation, though the gains are specific to the ABAW challenge setting.
AffectFlow-DINO tackles multi-task affect estimation from static face images by replacing deterministic predictions with a conditional rectified-flow head for uncertainty-aware outputs. It achieves a multi-task performance score of 1.177, significantly surpassing the challenge baseline of 0.45, with notable gains in valence-arousal estimation (CCC-V +0.058) and rare class recovery (Fear from 3.8% to 33.1%).
We present \textbf{AffectFlow-DINO}, a multi-task learning system for the 11th ABAW challenge that extends a standard deterministic architecture with a conditional rectified-flow head to model the inherent ambiguity of in-the-wild facial behavior. Instead of predicting a single affect estimate, the model learns a conditional generative distribution, enabling uncertainty-aware one-to-many predictions through Monte Carlo sampling. The system jointly estimates continuous valence-arousal, classifies eight facial expressions, and detects twelve Action Units from static face images. Built on a frozen DINOv3 ViT-S/16 backbone, extensive ablation studies show that rectified-flow decoding consistently improves deterministic prediction, particularly for valence-arousal estimation (CCC-V $+0.058$). We further show that post-hoc threshold calibration effectively recovers performance on severely imbalanced rare classes (e.g., Fear: $3.8\% \rightarrow 33.1\%$) without retraining. Combined with backbone fine-tuning and flow retuning, the final model achieves $\mathbf{P_{MTL}=1.177}$, substantially outperforming the official challenge baseline of $P_{MTL}=0.45$.