Papers › From Poses to Identity: Training-Free Person Re-Identification via Feature Centralization

From Poses to Identity: Training-Free Person Re-Identification via Feature Centralization

2 Mar 2025CVPR 2025 1arXiv:2503.00938archive 2025-07-28

Chao Yuan, Guiwei Zhang, Changxiao Ma, Tianyi Zhang, Guanglin Niu

Person re-identification (ReID) aims to extract accurate identity representation features. However, during feature extraction, individual samples are inevitably affected by noise (background, occlusions, and model limitations). Considering that features from the same identity follow a normal distribution around identity centers after training, we propose a Training-Free Feature Centralization ReID framework (Pose2ID) by aggregating the same identity features to reduce individual noise and enhance the stability of identity representation, which preserves the feature's original distribution for following strategies such as re-ranking. Specifically, to obtain samples of the same identity, we introduce two components:Identity-Guided Pedestrian Generation: by leveraging identity features to guide the generation process, we obtain high-quality images with diverse poses, ensuring identity consistency even in complex scenarios such as infrared, and occlusion.Neighbor Feature Centralization: it explores each sample's potential positive samples from its neighborhood. Experiments demonstrate that our generative model exhibits strong generalization capabilities and maintains high identity consistency. With the Feature Centralization framework, we achieve impressive performance even with an ImageNet pre-trained model without ReID training, reaching mAP/Rank-1 of 52.81/78.92 on Market1501. Moreover, our method sets new state-of-the-art results across standard, cross-modality, and occluded ReID tasks, showcasing strong adaptability.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

yuanc3/Pose2ID officialmentioned on GitHubpytorch report
yuanc3/dmon-aro mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Cross-Modal Person Re-IdentificationPerson Re-IdentificationRe-Ranking

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Cross-Modal Person Re-Identification SYSU-MM01 SAAI+Pose2ID mAP (All-search & Single-shot) 76.44 #1 of 1 Archive leaderboard report
Cross-Modal Person Re-Identification SYSU-MM01 SAAI+Pose2ID rank1 79.33 #1 of 1 Archive leaderboard report
Person Re-Identification Market-1501 CLIP-ReID+Pose2ID (no RK) Rank-1 97.3 #3 of 135 Archive leaderboard report
Person Re-Identification Market-1501 CLIP-ReID+Pose2ID (no RK) mAP 94.9 #3 of 135 Archive leaderboard report
Person Re-Identification Market-1501 TransReID+Pose2ID (no RK) Rank-1 95.52 #52 of 135 Archive leaderboard report
Person Re-Identification Market-1501 TransReID+Pose2ID (no RK) mAP 93.01 #52 of 135 Archive leaderboard report
Person Re-Identification Occluded REID KPR + Pose2ID (no RK) Rank-1 91.00 #1 of 5 Archive leaderboard report
Person Re-Identification Occluded REID KPR + Pose2ID (no RK) mAP 89.34 #1 of 5 Archive leaderboard report
Person Re-Identification Occluded REID BPBreID + Pose2ID (no RK) Rank-1 89.10 #2 of 5 Archive leaderboard report
Person Re-Identification Occluded REID BPBreID + Pose2ID (no RK) mAP 86.05 #2 of 5 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections