AI Paper Atlas

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

Multi-Centroid Representation Network for Domain Adaptive Person Re-ID

Yuhang Wu, Tengteng Huang, Haotian Yao, Chi Zhang, Yuanjie Shao, Chuchu Han, Changxin Gao, Nong Sang

AAAI 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Supervised person re-identification · 64 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
replace uni-centroid cluster representation with multi-centroid memory (ablation)SpCL baseline · Duke→Market, Market→Duke, Duke→MSMT, Market→MSMTmAP: 73.1 -> 81.5; 62.1 -> 70.8; 19.1 -> 30.2; 18.2 -> 28.1Improved
add DSCL, restricting contrastive comparisons to the query's domain (ablation)MCM · Duke→Market, Market→Duke, Duke→MSMT, Market→MSMTmAP: 81.5 -> 82.4; 70.8 -> 71.4; 30.2 -> 31.1; 28.1 -> 29.7Improved
add SONI synthetic hard negatives (ablation)MCM · Duke→Market, Market→Duke, Duke→MSMT, Market→MSMTmAP: 81.5 -> 82.9; 70.8 -> 70.8; 30.2 -> 33.7; 28.1 -> 32.3Improved
choose the most-similar positive centroid rather than the moderate one (ablation)MCM · Duke→MarketmAP: 81.5 -> 56.3Hurt
choose the least-similar positive centroid rather than the moderate one (ablation)MCM · Duke→Market, Market→DukemAP: 81.5 -> 79.9; 70.8 -> 69.4Hurt
use all negative centroids instead of their per-cluster mean (ablation)MCM with moderate positive selection · Duke→Market, Market→DukemAP: 81.5 -> 81.0; 70.8 -> 70.0Hurt
use mean negative centroids rather than all negative centroids (ablation)MCM with moderate positive selection · Duke→Market, Market→DukemAP: 81.0 -> 81.5; 70.0 -> 70.8Improved
increase centroid count per class from 2 to 4 (ablation)MCRN · Duke→MarketmAP: 75.7 -> 83.8Improved
increase centroid count per class above 4 (ablation)MCRN · Duke→MarketmAP: 83.8; 83.5; 83.5No clear effect
use QNNI interpolation between the query and a nearest negative instead of SONI (ablation)MCRN · Duke→Market, Market→DukemAP: 83.8 vs 76.5; 71.5 vs 66.1Hurt
use RNNI interpolation between two randomly selected top-nearest negatives instead of SONI (ablation)MCRN · Duke→Market, Market→DukemAP: 83.8 vs 83.4; 71.5 vs 70.4Hurt
add MCM on source-domain data (ablation)baseline · Duke→MarketmAP: +1.8%Improved

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