Event-driven Re-Id: A New Benchmark and Method Towards Privacy-Preserving Person Re-Identification
Abstract
The large-scale use of surveillance cameras in public spaces raised severe concerns about an individual privacy breach. Introducing privacy and security in video surveillance systems, primarily in person re-identification (re-id), is quite challenging. Event cameras are novel sensors, which only respond to brightness changes in the scene. This characteristic makes event-based vision sensors viable for privacy-preserving in video surveillance. Integrating privacy into the person re-id; this work investigates the possibility of performing person re-id with the event-camera network for the first time. We transform the asynchronous events stream generated by an event camera into synchronous image-like representations to leverage deep learning models and then evaluate how complex the re-id problem is with this new sensor modality. Interestingly, such event-based representations contain meaningful spatial details which are very similar to standard edges and contours. We use two different representations, image-like representation and their transformation to polar coordinates (which carry more distinct edge patterns). Finally, we train a person re-id model on such images to demonstrate the feasibility of performing event-driven re-id. We evaluate the performance of our approach and produce baseline results on two synthetic datasets (generated from publicly available datasets, SAIVT and DukeMTMC-reid).
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Ahmad et al. [7] introduced a novel event-based person ReId system.
“Ahmad et al. [7] introduced a novel event-based person ReId system to tackle this issue.”
From Event Anonymization: Privacy-Preserving Person Re-Identification and Pose Estimation in Event-Based Vision (2024) · cites this paper as [7]
The research community lacks a dataset captured with real event cameras that is appropriate for benchmarking person ReId methods.
“Yet, the research community lacks a dataset captured with real event cameras, which are also appropriate for benchmarking person ReId methods.”
From Event Anonymization: Privacy-Preserving Person Re-Identification and Pose Estimation in Event-Based Vision (2024) · cites this paper as [7]
Event cameras capture only brightness changes in the scene, not detailed visual images of people, and are often considered privacy-preserving because they eliminate detailed visual private data such as face details.
“Event cameras are often considered privacy-preserving [7], [35] because they inherently eliminate detailed visual private data such as face details.”
From Event Anonymization: Privacy-Preserving Person Re-Identification and Pose Estimation in Event-Based Vision (2024) · cites this paper as [7]
Ahmad et al. [7] demonstrated that event frames primarily provide edge and texture information, potentially useful for ReId purposes.
“Ahmad et al. [7] demonstrated that event frames primarily provide edge and texture information, potentially useful for ReId purposes.”
From Event Anonymization: Privacy-Preserving Person Re-Identification and Pose Estimation in Event-Based Vision (2024) · cites this paper as [7]
Even after event-stream anonymization, person ReId performance is still better than Ed-ReId, although there is a decrease when applying privacy.
“Results in Table 5 illustrate that even after event-stream anonymization with our proposed network, person ReId performance is still better than Ed-ReId, although we pay a reasonable decrease in the score when applying privacy (i.e., the anonymization module).”
From Event Anonymization: Privacy-Preserving Person Re-Identification and Pose Estimation in Event-Based Vision (2024) · cites this paper as [7]
The citation lists haven’t been fetched yet
References 45 · Cited by 54
- Authors
- Ahmad, S., Scarpellini, G., Morerio, P., & Bue, A. D.
- Affiliations
- Italian Institute of Technology (IT)
- Venue
- 2022 IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW)
- Publisher
- IEEE
- Published
- 2022-01
- Pages
- 459-468
- OpenAlex
- W4213277438
- Cited by
- 54
- References
- 45
- Citations by year
- 2026: 11 · 2025: 13 · 2024: 17 · 2023: 10 · 2022: 3
- FWCI
- 3.90
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Surveillance and Tracking Methods
- Physical Sciences › Engineering › Biomedical Engineering › Gait Recognition and Analysis
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Keywords
- person re-identification, event cameras, event-based vision, privacy-preserving monitoring, video surveillance, polar representation, synthetic benchmark, DukeMTMC-reID
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
