Globally Optimal Solution to Multi-object Tracking with Merged Measurements

lastmod 2026-11-10created 2017-03-022 edits

Globally Optimal Solution to Multi-object Tracking with Merged Measurements is an ICCV 2011 paper that incorporates cases where objects in multi-object tracking come close together and clump into a single detection into a flow network model, linking trajectories before and after occlusion without breaking them, in a globally optimal way. This page archives the author's project page and PDF; a separate page exists for the same paper with only a different title spelling.

Documentation

Globally Optimal Solution to Multi-object Tracking with Merged Measurements

http://www.robots.ox.ac.uk/~joao/merged/