AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

HierUQ: Hierarchical Uncertainty Quantification with Adaptive Granularity Reconciliation for Degraded Image Classification

Yang Chu, Xiaomeng Yang, Keli Deng, Yuntao Qian

CVPR 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Hierarchical Uncertainty Quantification (HUQ) (ablation)ViT baseline · HRSC-Deg (2-level)level accuracy: 68.08 -> 71.20Improved
add HUQ (ablation)ViT baseline · CUB-Deg (3-level)species accuracy: +3.69%Improved
add Adaptive Granularity Reconciliation (BiLT) (ablation)ViT baseline · HRSC-Deg (2-level)fine accuracy: +8.31%Improved
add Adaptive Granularity Reconciliation (BiLT) (ablation)ViT baseline · CUB-Deg (3-level)species accuracy: +1.37%Improved
add Confidence-Aware Hierarchical Fallback (CAHF) (ablation)ViT baseline · HRSC-Deg (2-level)ISDL: 58.63 -> 67.26Improved
combine HUQ, AGR and CAHF (ablation)ViT baseline · HRSC-Deg (2-level)ISDL: 58.63 -> 85.45Improved
combine HUQ, AGR and CAHF (ablation)ViT baseline · CUB-Deg (3-level)ISDL: 64.97 -> 80.91Improved
add bilinear fusion to HierUQ-C (ablation)HierUQ · HRSC-Deg (2-level)ISDL: 76.27 -> 85.45Improved
add bilinear fusion to HierUQ-C (ablation)HierUQ · CUB-Deg (3-level)species accuracy: 85.06 -> 85.73Improved
add multi-level joint optimization (MLJO) (ablation)HierUQ training · CUB-Deg (3-level)epochs to converge: 38 -> 18 epochsImproved
add HUQ to training (ablation)HierUQ · CUB-Deg (3-level)uncertainty improvement score: above 3.0 vs approaching 1.0Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy