Object Tracking
Object tracking, the task of identifying and following objects in image or video sequences, aims to accurately estimate object locations and trajectories over time. Current research emphasizes improving robustness in challenging conditions (low light, underwater, camouflage) and expanding capabilities to handle multiple objects, open-vocabulary categories, and non-linear motion, often leveraging transformer networks, generative models, and multimodal fusion (e.g., combining camera and radar data). These advancements have significant implications for various fields, including autonomous driving, robotics, surveillance, and scientific studies of animal behavior, by enabling more reliable and efficient automated systems.
Papers
Addressing single object tracking in satellite imagery through prompt-engineered solutions
Athena Psalta, Vasileios Tsironis, Andreas El Saer, Konstantinos Karantzalos
Learning Motion Blur Robust Vision Transformers with Dynamic Early Exit for Real-Time UAV Tracking
You Wu, Xucheng Wang, Dan Zeng, Hengzhou Ye, Xiaolan Xie, Qijun Zhao, Shuiwang Li
Reliable Object Tracking by Multimodal Hybrid Feature Extraction and Transformer-Based Fusion
Hongze Sun, Rui Liu, Wuque Cai, Jun Wang, Yue Wang, Huajin Tang, Yan Cui, Dezhong Yao, Daqing Guo
XTrack: Multimodal Training Boosts RGB-X Video Object Trackers
Yuedong Tan, Zongwei Wu, Yuqian Fu, Zhuyun Zhou, Guolei Sun, Eduard Zamfi, Chao Ma, Danda Pani Paudel, Luc Van Gool, Radu Timofte