Weak-structure-aware visual object tracking with bottom-up and top-down context exploration
作者：
Highlights：
•
• A weak-structure-aware solution is proposed for part-based object tracking.
• Two sets of parts are combined complementarily to find the object location and scale adaptively.
• A bottom-up procedure is introduced to explore pixel-level context for consistent tracking.
• A top-down procedure is devised to explore object-level context for effective model support.
• The proposed method is tested in extensive experiments to demonstrate its excellent performance.
摘要
•A weak-structure-aware solution is proposed for part-based object tracking.•Two sets of parts are combined complementarily to find the object location and scale adaptively.•A bottom-up procedure is introduced to explore pixel-level context for consistent tracking.•A top-down procedure is devised to explore object-level context for effective model support.•The proposed method is tested in extensive experiments to demonstrate its excellent performance.
论文关键词：Visual tracking,Weak structure,Keypoint,Context exploration
论文评审过程：Received 22 December 2017, Revised 9 August 2018, Accepted 9 August 2018, Available online 13 August 2018, Version of Record 23 August 2018.
论文官网地址：https://doi.org/10.1016/j.image.2018.08.005
