Research map / Label-efficient and robust learning / Recognition and long-tail
Supervised person re-identification
Learns discriminative person features with similarity learning, part features and bias removal.
108 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Recognition and long-tail, keyword cluster “face · person identification · facial”
State History Benchmarks Tried Baselines Checklist Preprints Who Export
Where it stands
Few recent reports The tracked benchmarks have at most one report from 2025–26 that could beat the old record, too few to say whether they are still moving.
11 papers in 2023–24 and 14 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 0.9% to 1.1%.
Open questions the data points to
- Papers report Mancs on DukeMTMC-reID (mAP) anywhere from 71.8 to 82.3: comparisons against it depend on whose number you take.
- Papers report Baseline on MSMT17 (Rank-1) anywhere from 34 to 68.8: comparisons against it depend on whose number you take.
No longer reported by papers here: DukeMTMC-reID (mAP, last in 2024); Market-1501 (Rank-1, last in 2024); DukeMTMC-reID (Rank-1, last in 2024); MSMT17 (Rank-1, last in 2022).
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Person Re-Identification by Multi-Channel Parts-Based CNN With Improved Triplet Loss Function (CVPR 2016 · 1,300 citations)
- 2017 · Re-Ranking Person Re-Identification With k-Reciprocal Encoding (CVPR 2017 · 1,534 citations)
- 2018 · Harmonious Attention Network for Person Re-Identification (CVPR 2018 · 1,430 citations)
- 2019 · Joint Discriminative and Generative Learning for Person Re-Identification (CVPR 2019 · 869 citations)
- 2020 · Relation-Aware Global Attention for Person Re-Identification (CVPR 2020 · 655 citations)
- 2021 · Diverse Part Discovery: Occluded Person Re-Identification With Part-Aware Transformer (CVPR 2021 · 421 citations)
- 2022 · Part-Based Pseudo Label Refinement for Unsupervised Person Re-Identification (CVPR 2022 · 296 citations)
- 2023 · Good Is Bad: Causality Inspired Cloth-Debiasing for Cloth-Changing Person Re-Identification (CVPR 2023 · 133 citations)
- 2024 · Occluded Person Re-identification via Saliency-Guided Patch Transfer (AAAI 2024 · 34 citations)
- 2025 · SeCap: Self-Calibrating and Adaptive Prompts for Cross-view Person Re-Identification in Aerial-Ground Networks (CVPR 2025 · 16 citations)
- 2026 · View-Aware Semantic Alignment for Aerial-Ground Person Re-Identification (CVPR 2026 · 0 citations)
Earliest papers here
- Person Re-Identification by Multi-Channel Parts-Based CNN With Improved Triplet Loss Function (CVPR 2016 · 1,300 citations)
- Learning Deep Feature Representations With Domain Guided Dropout for Person Re-Identification (CVPR 2016 · 990 citations)
- Deep Relative Distance Learning: Tell the Difference Between Similar Vehicles (CVPR 2016 · 821 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
Market-1501 · mAP (higher is better)
41 papers, 2017–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2017 | 63.63 | Re-Ranking Person Re-Identification With k-Reciprocal Encoding (CVPR 2017 · 1,534 citations) |
| 2018 | 82.5 | Deep Group-Shuffling Random Walk for Person Re-Identification (CVPR 2018 · 146 citations) |
| 2019 | 92.38 | AANet: Attribute Attention Network for Person Re-Identifications (CVPR 2019 · 366 citations) |
| 2020 | 95.8 | Unity Style Transfer for Person Re-Identification (CVPR 2020 · 86 citations) |
DukeMTMC-reID · mAP (higher is better)
29 papers, 2018–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2018 | 69.5 | Group Consistent Similarity Learning via Deep CRF for Person Re-Identification (CVPR 2018 · 271 citations) |
| 2019 | 86.87 | AANet: Attribute Attention Network for Person Re-Identifications (CVPR 2019 · 366 citations) |
| 2020 | 91.82 | Viewpoint-Aware Loss with Angular Regularization for Person Re-Identification (AAAI 2020 · 100 citations) |
Market-1501 · Rank-1 (higher is better)
27 papers, 2017–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2017 | 77.11 | Re-Ranking Person Re-Identification With k-Reciprocal Encoding (CVPR 2017 · 1,534 citations) |
| 2018 | 86.94 | Features for Multi-Target Multi-Camera Tracking and Re-Identification (CVPR 2018 · 551 citations) |
| 2019 | 94.8 | Joint Discriminative and Generative Learning for Person Re-Identification (CVPR 2019 · 869 citations) |
| 2020 | 97.9 | Online Joint Multi-Metric Adaptation From Frequent Sharing-Subset Mining for Person Re-Identification (CVPR 2020 · 53 citations) |
DukeMTMC-reID · Rank-1 (higher is better)
24 papers, 2018–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2018 | 77.69 | Features for Multi-Target Multi-Camera Tracking and Re-Identification (CVPR 2018 · 551 citations) |
| 2019 | 87.7 | Re-Identification With Consistent Attentive Siamese Networks (CVPR 2019 · 280 citations) |
| 2020 | 93.85 | Viewpoint-Aware Loss with Angular Regularization for Person Re-Identification (AAAI 2020 · 100 citations) |
MSMT17 · mAP (higher is better)
16 papers, 2019–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 52.3 | Joint Discriminative and Generative Learning for Person Re-Identification (CVPR 2019 · 869 citations) |
| 2020 | 64.2 | Online Joint Multi-Metric Adaptation From Frequent Sharing-Subset Mining for Person Re-Identification (CVPR 2020 · 53 citations) |
| 2021 | 65.7 | Unsupervised Pre-Training for Person Re-Identification (CVPR 2021 · 206 citations) |
| 2024 | 74.1 | CA-Jaccard: Camera-aware Jaccard Distance for Person Re-identification (CVPR 2024 · 31 citations) |
MSMT17 · Rank-1 (higher is better)
10 papers, 2019–2022 · not reported since 2022
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 77.2 | Joint Discriminative and Generative Learning for Person Re-Identification (CVPR 2019 · 869 citations) |
| 2020 | 85.7 | Online Joint Multi-Metric Adaptation From Frequent Sharing-Subset Mining for Person Re-Identification (CVPR 2020 · 53 citations) |
What has been tried
95 papers here report 1,075 measured changes (a component added, removed or swapped, compared with a baseline): helped 79%, no clear effect 3%, hurt 16%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add OIoU to random SPT pairing | Occluded-Duke · Rank-1 | 65.3 -> 66.6 | Helped | Occluded Person Re-identification via Saliency-Guided Patch Transfer (AAAI 2024 · 34 citations) |
| Replace robust KRNNs with CKRNNs | Market1501 · mAP | 81.0 -> 82.4 | Helped | CA-Jaccard: Camera-aware Jaccard Distance for Person Re-identification (CVPR 2024 · 31 citations) |
| Remove the global loss, leaving only local loss | RAP, no cross att. · mAP | 74.26 -> 69.98 | Hurt | Attribute-Guided Pedestrian Retrieval: Bridging Person Re-ID with Internal Attribute Variability (CVPR 2024 · 24 citations) |
| Add ADM-based artificial occlusion | Occluded-Duke · Rank-1 | 59.7 -> 66.2; +6.5\% | Helped | Attention Disturbance and Dual-Path Constraint Network for Occluded Person Re-identification (AAAI 2024 · 23 citations) |
| Add VDT to the ViT baseline | LAGPeR A→G · Rank-1 | 38.67 -> 40.15 | Helped | SeCap: Self-Calibrating and Adaptive Prompts for Cross-view Person Re-Identification in Aerial-Ground Networks (CVPR 2025 · 16 citations) |
| multi-task pre-training (point cloud completion + SMPL shape) on LReID-sync | LReID · Rank-1 | 90.0 to 94.0 | Helped | LiDAR-based Person Re-identification (CVPR 2024 · 16 citations) |
| Add DALNet’s auxiliary modality and cross-modality circle loss | PKU-Sketch · Rank-1 | 67.8 -> 76.5 | Helped | Differentiable Auxiliary Learning for Sketch Re-Identification (AAAI 2024 · 11 citations) |
| Use expand_resize instead of random_paste for re-ID data unification | PRW · mAP | 46.49; 47.08 | Helped | Divide and Conquer: Hybrid Pre-training for Person Search (AAAI 2024 · 8 citations) |
| Add CDC to the baseline | Market-1501 · mAP | 82.8 to 86.1 | Helped | CDE-Learning: Camera Deviation Elimination Learning for Unsupervised Person Re-identification (AAAI 2025 · 6 citations) |
| Add DDA to the baseline | NightReID · Rank-1 | 65.0 -> 66.1 | Helped | NightReID: A Large-Scale Nighttime Person Re-Identification Benchmark (AAAI 2025 · 6 citations) |
| Use two feedback samples per query instead of one (K=2 rather than K=1; T=1) | Market1501 · mAP | 94.55 -> 94.74 | Helped | QI-IRA: Quantum-Inspired Interactive Ranking Aggregation for Person Re-identification (AAAI 2024 · 4 citations) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
| Baseline | Benchmark | Metric | Reported | Papers |
|---|---|---|---|---|
| Mancs | DukeMTMC-reID | mAP | 71.8 – 82.3 | 3 papers, e.g. CVPR 2019 and CVPR 2020 |
| Baseline | MSMT17 | Rank-1 | 34 – 68.8 | 2 papers, e.g. AAAI 2021 and CVPR 2019 |
| Mancs | DukeMTMC-reID | Rank@1 | 84.9 – 93.1 | 2 papers, e.g. CVPR 2019 and CVPR 2020 |
What reviewers will expect
Datasets and baselines named in the 6 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- TransReID · 2 of 6 papers
- CLIP-ReID · 2 of 6 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 25 papers; from OpenAlex and the papers' affiliations).
- Wuhan University · 4 papers
- Xidian University · 2 papers
- Xiamen University · 2 papers
- Tencent · 2 papers
- Dalian Maritime University · 2 papers
Use it
Download BibTeX of 40 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Margin-based face recognition losses (124 papers)
- Cross-modality and occlusion re-identification (82 papers)
- Fair and quality-aware face recognition (66 papers)
- Uncertainty-aware facial expression recognition (41 papers)
Elsewhere in the topic
- Multilingual language representation pretraining (507 papers)
- Contrastive representation theory (495 papers)
- Dense features and foundation priors (482 papers)
- Correspondence and causal representation learning (447 papers)
- Data augmentation and generalization bias (409 papers)
- Partial and multi-label learning (345 papers)
- Loss design and classifier theory (273 papers)
- Group-robust classification (251 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).