AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Do Different Tracking Tasks Require Different Appearance Models?

Zhongdao Wang, Hengshuang Zhao, Ya-Li Li, Shengjin Wang, Philip Torr, Luca Bertinetto

NeurIPS 2021 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Pretext-task visual pre-training · 32 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
Use UniTrack ImageNet appearance features instead of FairMOT's task-trained appearance modelMOT-16 tracking · MOT-16 testIDs: 1074 to 683Improved
Use a shared UniTrack appearance model instead of LightTrackPoseTrack · PoseTrack2018 valIDF1: +21.0 pointsImproved
Use UniTrack instead of LightTrackPoseTrack · PoseTrack2018 valMOTA: 1.3 pointsHurt
Use UniTrack ImageNet DCF head instead of XCorr head (ablation)UniTrack box propagation · OTB-2015AUC: 55.5 -> 61.8Improved
Use UniTrack DCF head instead of SiamFCSingle-object tracking · OTB-2015AUC: 3.6 AUC pointsImproved
Use UniTrack ImageNet mask propagation instead of SiamMaskVideo object segmentation · DAVIS-2017J-mean: 4.1 J-mean pointsImproved
Use RSM instead of global-pooled features (ablation)UniTrack association · MOT-16 train, DPM detections, motion cuesIDF1: 34.4 -> 39.6Improved
Use RSM instead of supervised ReID features (ablation)UniTrack association · MOT-16 train, FRCNN detections, appearance onlyIDF1: 47.0 -> 51.5Improved
Use a shared UniTrack representation instead of COSTAMOTS tracking · MOTS testIDF1: 70.3 vs 68.2Hurt
Use a shared UniTrack representation instead of COSTAMOTS tracking · MOTS testIDs: 421 vs 342Improved
Use a shared UniTrack representation instead of KeyTrackPoseTrack · PoseTrack2018 valMOTA: 66.6 vs 63.3Hurt
Use UniTrack VFS instead of ImageNet featuresVideo object segmentation · DAVIS-2017J-mean: 58.4 -> 62.8Improved

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