AttentionTrack: Multiple Object Tracking in Traffic Scenarios Using Features Attention

Multiple object tracking (MOT) is becoming increasingly significant for autonomous driving and intelligent transportation systems. However, traditional MOT methods cannot track the objects accurately and robustly due to the lack of effective feature extraction and data association in complex traffic scenarios. In this paper, the authors propose a novel joint detection and tracking method AttentionTrack by introducing multiple features attention. Firstly, they design a self-motivated feature extraction attention network (FEAN) to adaptively produce effective decoupled features for detection and tracking tasks in different scenarios. Secondly, they build a spatial-temporal data association (STDA) framework to achieve more accurate and robust tracking by considering the historical features of trajectory through different times. Moreover, they conduct comprehensive experiments on the KITTI, UA-DETRAC and MOT17 benchmarks, and the results show that our approach achieves competitive performance compared with the state-of-the-art (SOTA) trackers.

• Record URL:
  • https://doi.org/10.1109/TITS.2023.3315222
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/oclc/41297384
• Supplemental Notes:
  • Copyright © 2024, IEEE.
• Authors:
  • Zhang, Chuang
  • Zheng, Sifa
  • Wu, Haoran
  • Gu, Ziqing
  • Sun, Wenchao
  • Yang, Lei
• Publication Date: 2024-2

Language

• English

Media Info

• Media Type: Web
• Features: References;
• Pagination: pp 1661-1674
• Serial:
  • IEEE Transactions on Intelligent Transportation Systems
  • Volume: 25
  • Issue Number: 2
  • Publisher: Institute of Electrical and Electronics Engineers (IEEE)
  • ISSN: 1524-9050
  • Serial URL: http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=6979

Subject/Index Terms

• TRT Terms: Attention; Autonomous vehicles; Detection and identification systems; Tracking systems
• Subject Areas: Highways; Vehicles and Equipment;

Filing Info

• Accession Number: 01919820
• Record Type: Publication
• Files: TRIS
• Created Date: May 28 2024 10:44AM