English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Multi-Camera Multi-Object Tracking is currently drawing attention in thecomputer vision field due to its superior performance in real-worldapplications such as video surveillance in crowded scenes or in wide spaces. Inthis work, we propose a mathematically elegant multi-camera multiple objecttracking approach based on a spatial-temporal lifted multicut formulation. Ourmodel utilizes state-of-the-art tracklets produced by single-camera trackers asproposals. As these tracklets may contain ID-Switch errors, we refine themthrough a novel pre-clustering obtained from 3D geometry projections. As aresult, we derive a better tracking graph without ID switches and more preciseaffinity costs for the data association phase. Tracklets are then matched tomulti-camera trajectories by solving a global lifted multicut formulation thatincorporates short and long-range temporal interactions on tracklets located inthe same camera as well as inter-camera ones. Experimental results on theWildTrack dataset yield near-perfect performance, outperformingstate-of-the-art trackers on Campus while being on par on the PETS-09 dataset.

LMGP: Lifted Multicut Meets Geometry Projections for Multi-Camera Multi-Object Tracking