Combining complementary trackers for enhanced long-term visual object tracking
作者：
Highlights：
• Fusing the characteristics of complementary trackers in long-term visual object tracking.
• Combination achieved through an evaluation, decision and correction strategy.
• Tracker evaluation implemented as an online learned deep verification model.
• Increased tracking accuracy and robustness is achieved.
• State-of-the-art performance on the most popular long-term visual tracking benchmarks.
摘要
Highlights•Fusing the characteristics of complementary trackers in long-term visual object tracking.•Combination achieved through an evaluation, decision and correction strategy.•Tracker evaluation implemented as an online learned deep verification model.•Increased tracking accuracy and robustness is achieved.•State-of-the-art performance on the most popular long-term visual tracking benchmarks.
论文关键词：Video tracking,Visual object tracking,Long-term visual tracking,Deep learning
论文评审过程：Received 22 December 2021, Revised 29 March 2022, Accepted 4 April 2022, Available online 12 April 2022, Version of Record 25 April 2022.
论文官网地址：https://doi.org/10.1016/j.imavis.2022.104448
