Ensemble Tracking Based on Diverse Collaborative Framework With Multi-Cue Dynamic Fusion

• Published in: IEEE transactions on multimedia Vol. 22; no. 10; pp. 2698 - 2710
• Main Authors: Han, Yamin, Zhang, Peng, Zhuo, Tao, Huang, Wei, Zha, Yufei, Zhang, Yanning
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.10.2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptation models; Artificial neural networks; Collaboration; collaborative tracking framework; Correlation; Ensembling structure; heuristic frequency; Machine learning; multi-cue dynamic fusion; Robustness; Target tracking; Tracking; Tracking loops; Training

Tracking with deep neural networks has been verified to arrive at a new level accuracy in many challenging scenarios, but the tracking robustness has been still challenged by model singularity and self-learning loop mechanism. As a promising solution for the limitations, to ensemble diverse tracking...