APTrack shows that a single unified multi-modal tracker using equal modality modeling and learnable token interaction can beat both unified and task-specific trackers on RGB-T, RGB-D, and RGB-E benchmarks.
Visual prompt multi- modal tracking,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Adaptive Perception for Unified Visual Multi-modal Object Tracking
APTrack shows that a single unified multi-modal tracker using equal modality modeling and learnable token interaction can beat both unified and task-specific trackers on RGB-T, RGB-D, and RGB-E benchmarks.