UAPAR is the first evidential deep learning framework for pedestrian attribute recognition that estimates attribute-wise epistemic uncertainty via a region-aware module and uses uncertainty-guided curriculum learning to handle label noise, achieving competitive results on PA100K, PETA, RAPv1 and RAP
Wentao Bao, Qi Yu, and Yu Kong
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Uncertainty-Aware Pedestrian Attribute Recognition via Evidential Deep Learning
UAPAR is the first evidential deep learning framework for pedestrian attribute recognition that estimates attribute-wise epistemic uncertainty via a region-aware module and uses uncertainty-guided curriculum learning to handle label noise, achieving competitive results on PA100K, PETA, RAPv1 and RAP