# Person Re-identification in the Wild

arXiv: https://arxiv.org/abs/1604.02531 · HTML: https://mlsanity.com/paper/1604.02531/

## Key facts

- Cited by 157 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Apr 2016.

## Built on by (most cited first, 30 of 157)

- [ByteTrack: Multi-Object Tracking by Associating Every Detection Box](https://mlsanity.com/paper/2110.06864/) (2021)
- [Beyond Appearance: a Semantic Controllable Self-Supervised Learning Framework for Human-Centric Visual Tasks](https://mlsanity.com/paper/2303.17602/) (2023)
- [SoccerNet-Tracking: Multiple Object Tracking Dataset and Benchmark in Soccer Videos](https://mlsanity.com/paper/2204.06918/) (2022)
- [Multiple Object Tracking with Correlation Learning](https://mlsanity.com/paper/2104.03541/) (2021)
- [Learning Multi-Granular Hypergraphs for Video-Based Person Re-Identification](https://mlsanity.com/paper/2104.14913/) (2021)
- [NFormer: Robust Person Re-identification with Neighbor Transformer](https://mlsanity.com/paper/2204.09331/) (2022)
- [PLIP: Language-Image Pre-training for Person Representation Learning](https://mlsanity.com/paper/2305.08386/) (2023)
- [Anchor-Free Person Search](https://mlsanity.com/paper/2103.11617/) (2021)
- [TransCenter: Transformers with Dense Representations for Multiple-Object Tracking](https://mlsanity.com/paper/2103.15145/) (2021)
- [Large-Scale Pre-training for Person Re-identification with Noisy Labels](https://mlsanity.com/paper/2203.16533/) (2022)
- [Sequential End-to-end Network for Efficient Person Search](https://mlsanity.com/paper/2103.10148/) (2021)
- [Graph Sampling Based Deep Metric Learning for Generalizable Person Re-Identification](https://mlsanity.com/paper/2104.01546/) (2021)
- [DyGLIP: A Dynamic Graph Model with Link Prediction for Accurate Multi-Camera Multiple Object Tracking](https://mlsanity.com/paper/2106.06856/) (2021)
- [Cascade Transformers for End-to-End Person Search](https://mlsanity.com/paper/2203.09642/) (2022)
- [SMILEtrack: SiMIlarity LEarning for Occlusion-Aware Multiple Object Tracking](https://mlsanity.com/paper/2211.08824/) (2022)
- [PSTR: End-to-End One-Step Person Search With Transformers](https://mlsanity.com/paper/2204.03340/) (2022)
- [One More Check: Making "Fake Background" Be Tracked Again](https://mlsanity.com/paper/2104.09441/) (2021)
- [Deep learning-based person re-identification methods: A survey and outlook of recent works](https://mlsanity.com/paper/2110.04764/) (2021)
- [Context Understanding in Computer Vision: A Survey](https://mlsanity.com/paper/2302.05011/) (2023)
- [Transformer for Object Re-Identification: A Survey](https://mlsanity.com/paper/2401.06960/) (2024)
- [Referring to Any Person](https://mlsanity.com/paper/2503.08507/) (2025)
- [Multi-Prompts Learning with Cross-Modal Alignment for Attribute-based Person Re-Identification](https://mlsanity.com/paper/2312.16797/) (2023)
- [Decoupled and Memory-Reinforced Networks: Towards Effective Feature Learning for One-Step Person Search](https://mlsanity.com/paper/2102.10795/) (2021)
- [Erasing, Transforming, and Noising Defense Network for Occluded Person Re-Identification](https://mlsanity.com/paper/2307.07187/) (2023)
- [MSO: Multi-Feature Space Joint Optimization Network for RGB-Infrared Person Re-Identification](https://mlsanity.com/paper/2110.11264/) (2021)
- [Delving into the Trajectory Long-tail Distribution for Muti-object Tracking](https://mlsanity.com/paper/2403.04700/) (2024)
- [Quality-aware Part Models for Occluded Person Re-identification](https://mlsanity.com/paper/2201.00107/) (2022)
- [Adversarial Patch Attacks and Defences in Vision-Based Tasks: A Survey](https://mlsanity.com/paper/2206.08304/) (2022)
- [NumbOD: A Spatial-Frequency Fusion Attack Against Object Detectors](https://mlsanity.com/paper/2412.16955/) (2024)
- [GOAL: Global-local Object Alignment Learning](https://mlsanity.com/paper/2503.17782/) (2025)

## Abstract (arXiv)

We present a novel large-scale dataset and comprehensive baselines for end-to-end pedestrian detection and person recognition in raw video frames. Our baselines address three issues: the performance of various combinations of detectors and recognizers, mechanisms for pedestrian detection to help improve overall re-identification accuracy and assessing the effectiveness of different detectors for re-identification. We make three distinct contributions. First, a new dataset, PRW, is introduced to evaluate Person Re-identification in the Wild, using videos acquired through six synchronized cameras. It contains 932 identities and 11,816 frames in which pedestrians are annotated with their bounding box positions and identities. Extensive benchmarking results are presented on this dataset. Second, we show that pedestrian detection aids re-identification through two simple yet effective improvements: a discriminatively trained ID-discriminative Embedding (IDE) in the person subspace using convolutional neural network (CNN) features and a Confidence Weighted Similarity (CWS) metric that incorporates detection scores into similarity measurement. Third, we derive insights in evaluating detector performance for the particular scenario of accurate person re-identification.