
OneTracker unifies various tracking tasks by pretraining a Foundation Tracker on RGB datasets and adapting it to downstream RGB+X tasks using prompt-tuning techniques.


coremsg

onetracker-unifying-visual-object-tracking-with-foundation-models-and-efficient-tuning


OneTracker: Unifying Visual Object Tracking with Foundation Models and Efficient Tuning


title_rewrite


ACTrackは、Transformerのパラメータを凍結し、軽量な条件付きネットを追加して、追跡中の空間相関と時間変動をモデル化する新しい追跡フレームワークです。


actrack-adding-spatio-temporal-condition-for-visual-object-tracking


ACTrack: Adding Spatio-Temporal Condition for Visual Object Tracking



OneTracker unifies RGB and RGB+X tracking tasks efficiently, achieving state-of-the-art performance by pretraining a Foundation Tracker and adapting it to downstream tasks using prompt-tuning techniques.


unifying-visual-object-tracking-with-onetracker


Unifying Visual Object Tracking with OneTracker
