Uncertainty-Aware Pedestrian Attribute Recognition via Evidential Deep Learning
This work addresses the need for reliable uncertainty estimation in pedestrian attribute recognition, a critical task for surveillance and autonomous driving, by proposing a novel method that outperforms deterministic approaches on low-quality samples.
UAPAR introduces the first evidential deep learning framework for pedestrian attribute recognition, achieving competitive or superior performance on PA100K, PETA, RAPv1, and RAPv2 datasets while effectively identifying unreliable predictions to enhance robustness.
We propose UAPAR, an Uncertainty-Aware Pedestrian Attribute Recognition framework. To the best of our knowledge, this is the first EDL-based uncertainty-aware framework for pedestrian attribute recognition (PAR). Unlike conventional deterministic methods, which fail to assess prediction reliability on low-quality samples, UAPAR effectively identifies unreliable predictions and thus enhances system robustness in complex real-world scenarios. To achieve this, UAPAR incorporates Evidential Deep Learning (EDL) into a CLIP-based architecture. Specifically, a Region-Aware Evidence Reasoning module employs cross-attention and spatial prior masks to capture fine-grained local features, which are further processed by an evidence head to estimate attribute-wise epistemic uncertainty. To further enhance training robustness, we develop an uncertainty-guided dual-stage curriculum learning strategy to alleviate the adverse effects of severe label noise during training. Extensive experiments on the PA100K, PETA, RAPv1, and RAPv2 datasets demonstrate that UAPAR achieves competitive or superior performance. Furthermore, qualitative results confirm that the proposed framework generates uncertainty estimates that are predictive of challenging or erroneous samples.