Reproduce “TAQ-MOTR: Tracking-Aware Detection Query Formation for End-to-End Multiple Object Tracking” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hanyang Liu, Hong Zhang, Ding Yuan, Yang Han, Yifan Yang
TAQ-MOTR forms detection queries from current-frame proposals and existing track states before they enter the shared decoder. Its Hybrid Query Generator combines image-based proposals with track memory and learnable detection priors. On DanceTrack test, it reports 72.0 HOTA and ranks highest in HOTA among compared tracking-by-propagation methods on all three benchmarks. This benchmark comparison does not isolate the mechanism.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.