Papers › ReMix: Training Generalized Person Re-identification on a Mixture of Data

ReMix: Training Generalized Person Re-identification on a Mixture of Data

29 Oct 2024arXiv:2410.21938archive 2025-07-28

Timur Mamedov, Anton Konushin, Vadim Konushin

Modern person re-identification (Re-ID) methods have a weak generalization ability and experience a major accuracy drop when capturing environments change. This is because existing multi-camera Re-ID datasets are limited in size and diversity, since such data is difficult to obtain. At the same time, enormous volumes of unlabeled single-camera records are available. Such data can be easily collected, and therefore, it is more diverse. Currently, single-camera data is used only for self-supervised pre-training of Re-ID methods. However, the diversity of single-camera data is suppressed by fine-tuning on limited multi-camera data after pre-training. In this paper, we propose ReMix, a generalized Re-ID method jointly trained on a mixture of limited labeled multi-camera and large unlabeled single-camera data. Effective training of our method is achieved through a novel data sampling strategy and new loss functions that are adapted for joint use with both types of data. Experiments show that ReMix has a high generalization ability and outperforms state-of-the-art methods in generalizable person Re-ID. To the best of our knowledge, this is the first work that explores joint training on a mixture of multi-camera and single-camera data in person Re-ID.

PaperPDF

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

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

Tasks

DiversityGeneralizable Person Re-identificationPerson Re-Identification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Generalizable Person Re-identification CUHK03-NP (detected) ReMix MSMT17->Rank-1 27.3 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification CUHK03-NP (detected) ReMix MSMT17->mAP 27.4 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification CUHK03-NP (detected) ReMix MSMT17-All->Rank-1 37.7 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification CUHK03-NP (detected) ReMix MSMT17-All->mAP 37.2 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification CUHK03-NP (detected) ReMix RandPerson->Rank-1 19.3 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification CUHK03-NP (detected) ReMix RandPerson->mAP 18.4 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix MSMT17->Rank1 71.6 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix MSMT17->mAP 52.8 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix MSMT17-All->Rank-1 77.6 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix MSMT17-All->mAP 61.6 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix Market-1501->Rank1 58.4 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix Market-1501->mAP 38.8 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix RandPerson->Rank1 63.2 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification DukeMTMC-reID ReMix RandPerson->mAP 42.8 #2 of 4 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix DukeMTMC-reID->Rank1 71.3 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix DukeMTMC-reID->mAP 43.0 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix MSMT17->Rank-1 78.2 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix MSMT17->mAP 52.4 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix MSMT17-All->Rank-1 84.0 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix MSMT17-All->mAP 61.0 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix RandPerson->Rank-1 72.7 #3 of 5 Archive leaderboard report
Generalizable Person Re-identification Market-1501 ReMix RandPerson->mAP 45.4 #3 of 5 Archive leaderboard report
Person Re-Identification DukeMTMC-reID ReMix Rank-1 89.6 #45 of 94 Archive leaderboard report
Person Re-Identification DukeMTMC-reID ReMix mAP 79.8 #45 of 94 Archive leaderboard report
Person Re-Identification MSMT17 ReMix Rank-1 84.8 #25 of 43 Archive leaderboard report
Person Re-Identification MSMT17 ReMix mAP 63.9 #25 of 43 Archive leaderboard report
Person Re-Identification Market-1501 ReMix Rank-1 96.2 #22 of 135 Archive leaderboard report
Person Re-Identification Market-1501 ReMix mAP 89.8 #22 of 135 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