Study tests multi-view strategies for real-world luggage re-identification
Original title: Luggage re-identification: A study of strategies towards a real-world application
An end-to-end pipeline detects and re-identifies luggage across video streams despite occlusion, clutter and crowds.
Luggage re-identification links the same bag across disjoint cameras for forensics, security, border management and lost-luggage recovery. Existing work mostly matches cropped images under clean conditions, which fails under occlusion, clutter and crowds. The paper studies an end-to-end pipeline from multiple video streams and evaluates view and shot selection and merging, plus training across environment datasets, on a real-world benchmark.
Unconfirmed: The abstract does not report quantitative results for the variants.
- Target settings include airports, train stations and border crossings.
- The study focuses on combining multiple shots and views rather than single crops.
- Variants are tested for overall re-identification performance in practical applications.
- Training on datasets from different environments is part of the evaluation.
From the original article
1 min read at arxiv.orgLuggage re-identification (ReID) is a critical task in video surveillance for forensics, security or border management (suspicious or abandoned luggage recognition), as well as luggage management (lost luggage recovery). It enables the association of observations of the same luggage across disjoint camera views in complex environments such as airports, train stations and border crossing.
By Quentin Dhôte · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.