AI Paper Atlas

Research map / Detection and segmentation / Tracking

UniTrack: Differentiable Graph Representation Learning for Multi-Object Tracking

Bishoy Galoaa, Xiangyu Bai, Utsav Nandi, Sai Siddhartha Vivek Dhir Rangoju, Somaieh Amraee, Sarah Ostadabbas

ICLR 2026 · Detection and segmentation › Tracking · approach: Memory and correlation-based MOT · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add UniTrack loss to GTRGTR · MOT17ID switches: 1445 -> 951Improved
add UniTrack loss to GTRGTR · MOT17MOTA / IDF1 / HOTA: +3.8% MOTA, +3.3% IDF1, +8.8% HOTAImproved
add UniTrack loss to TrackformerTrackformer · MOT17False positives: 1965 -> 1039Improved
add UniTrack loss to TrackformerTrackformer · MOT17False negatives: 21,893 -> 16,667Improved
add UniTrack loss to GTRGTR · SportsMOTMOTA / IDF1 / HOTA: +9.7% MOTA, +12.3% IDF1, +11.7% HOTAImproved
add UniTrack loss to GTRGTR · SportsMOTID switches: 2364 -> 1092Improved
add UniTrack loss to ByteTrackByteTrack · SportsMOTID switches: 3267 -> 2234Improved
add UniTrack loss to GTRGTR · DanceTrackHOTA: 43.7 -> 50.2Improved
add UniTrack loss to GTRGTR · DanceTrackIDF1: 45.9 -> 48.5Improved
add the spatial component (ablation)UT-Trackformer · MOT17MOTA / IDF1: without it MOTA drops 1.9% and IDF1 falls 1.2%Improved
add the temporal component (ablation)UT-Trackformer · MOT17HOTA: without it HOTA drops 6.2%Improved
add the temporal component (ablation)UT-Trackformer · MOT17MOTA: removing it raises MOTA 2.1% and ID switches 288 -> 380Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy