.png){kind=icon}
Return
MCTrack: Multi-cue spatio-temporal object tracking
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2025.129984.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Decouple the object representations into initial, dynamic, and historical cues. • HCAM efficiently models historical cue without multi-frame encoding costs. • Spatio-temporal decoding head fuses historical query to enhance object features. • Dual-criterion with tracking and temporal scores ensures reliable cue updates. • Achieves state-of-the-art performance across multiple tracking benchmarks.
Journal
IF:
7.5
Papers:
3.0W
Citations:
10.2W
Organization
.png){kind=icon}
