AI Paper Atlas

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

Hierarchical Feature Embedding for Attribute Recognition

Jie Yang, Jiarou Fan, Yiru Wang, Yige Wang, Weihao Gan, Lin Liu, Wei Wu

CVPR 2020 · Label-efficient and robust learning › Recognition and long-tail · approach: Margin-based face recognition losses · 54 citations (OpenAlex)

Check 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
Use hierarchical attribute and ID feature embedding with HFE lossAttribute recognition · Market 1501 attribute datasetmean accuracy: 91.27 -> 92.90Improved
Use hierarchical attribute and ID feature embedding with HFE lossAttribute recognition · Duke attribute datasetmean accuracy: 90.35 -> 91.77Improved
Use HFE for face attribute recognitionDeepID2 · CelebAaccuracy: 91.81 -> 92.17Improved
Add inter-class triplet loss to CE (ablation)Attribute recognition · Market 1501 attribute datasetmean accuracy: 88.82 -> 90.83Improved
Add intra-class ID triplet loss to CE (ablation)Attribute recognition · Market 1501 attribute datasetmean accuracy: 88.82 -> 91.27Improved
Combine inter-class and intra-class triplet losses (ablation)Attribute recognition · Market 1501 attribute datasetmean accuracy: 90.83 -> 92.44Improved
Add absolute boundary regularization (ablation)Inter-class and intra-class triplet losses · Market 1501 attribute datasetmean accuracy: 92.44 -> 92.73Improved
Replace fixed HFE loss weight with dynamic weight (ablation)HFE · Market 1501 attribute datasetmean accuracy: 92.73 -> 92.90Improved
Replace intra-class triplet loss with pairwise ID compactness loss (ablation)Inter-class and intra-class feature embedding · Market 1501 attribute datasetmean accuracy: 92.44 -> 92.30Hurt

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