AI Paper Atlas

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

Spatial and Semantic Consistency Regularizations for Pedestrian Attribute Recognition

Jian Jia, Xiaotang Chen, Kaiqi Huang

ICCV 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Margin-based face recognition losses · 77 citations (OpenAlex)

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 spatial consistency regularization (SPAC) to the baseline (ablation)Baseline ResNet50 · PETAmA: 81.15 -> 84.08Improved
add spatial consistency regularization (SPAC) to the baseline (ablation)Baseline ResNet50 · PA100KmA: 78.53 -> 79.15Improved
add spatial consistency regularization (SPAC) to the baseline (ablation)Baseline ResNet50 · RAPmA: 76.09 -> 78.55Improved
add semantic consistency regularization (SEMC) alone to the baseline (ablation)Baseline ResNet50 · PA100KmA: 78.53 -> 78.63No clear effect
add semantic consistency regularization (SEMC) alone to the baseline (ablation)Baseline ResNet50 · PETAmA: 81.15 -> 82.34Improved
add SEMC on top of SPAC (ablation)SPAC model · PA100KmA: 79.15 -> 80.09Improved
add weighted loss to SPAC+SEMC (ablation)SPAC+SEMC model · PA100KmA: 80.09 -> 81.87Improved
add weighted loss to SPAC+SEMC (ablation)SPAC+SEMC model · RAPmA: 80.26 -> 82.77Improved
add weighted loss alone to the baseline (ablation)Baseline ResNet50 · RAPmA: 76.09 -> 79.46Improved
use full SPAC+SEMC+weighted loss versus baselineBaseline ResNet50 · PETAmA: 81.15 -> 86.52Improved
raise confidence threshold tau from 0 to 0.9 (ablation)SSC_soft · PA100KmA: 79.63 -> 81.87Improved
lower momentum coefficient alpha from 0.9 to 0.1 (ablation)SSC_soft · PA100KmA: 81.87 -> 80.89Hurt

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