DETrack: Depth information is predictable for tracking

Open Webpage

Weiyu Zhao, Yizhuo Jiang, Yan Gao, Jie Li, Xinbo Gao

Published: 2025, Last Modified: 01 Apr 2025Neurocomputing 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•We propose a depth information module with extra dimension features for 3D tracking.•We propose a data association method with depth information to mitigate ID switches.•The proposed framework suits most MOT methods and was verified on open datasets.