AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Adaptive Cross-Domain Learning for Generalizable Person Re-identification

Pengyi Zhang, Huanzhang Dou, Yunlong Yu, Xi Li

ECCV 2022 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Cross-domain few-shot with transformers · 60 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 dynamic branch to baseline (ablation)ResNet50-IBN · Protocol-2 averagemAP: 40.2 -> 42.8Improved
add static branch to baseline (ablation)ResNet50-IBN · Protocol-2 averagemAP: 40.2 -> 40.7No clear effect
add static branch to dynamic branch (ablation)ACL · Protocol-2 averagemAP: 42.8 -> 45.3Improved
use meta-parameter generator instead of standard MLP weight generator (ablation)ACL · Protocol-2 averagemAP: 43.6 -> 45.3Improved
use domain-aware softmax meta-weights instead of identity weights (ablation)ACL · Protocol-2 averagemAP: 43.4 -> 45.3Improved
add domain-aware cluster loss on meta-weights (ablation)ACL · Protocol-2 averagemAP: 44.2 -> 45.3Improved
add dynamic-static fusion module (ablation)ACL · Protocol-2 averagemAP: 43.5 -> 45.3Improved
add weighted sum to normalization in fusion (ablation)ACL · Protocol-2 averagemAP: 44.0 -> 44.7Improved
add refinery channel attention to fusion (ablation)ACL · Protocol-2 averagemAP: 44.7 -> 45.3Improved
embed CODE-Block in all stages instead of stage 4 only (ablation)ACL · Protocol-2 averagemAP: 43.6 -> 45.3Improved
embed CODE-Block in stage 4 instead of stage 1 (ablation)ACL · Protocol-2 averagemAP: 42.7 -> 43.6Improved
replace MetaBIN with ACLACL · Protocol-2 averageRank-1: 49.5 -> 59.0Improved

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