Abstract:Object tracking is an important research direction in the field of computer vision, with the task of continuously detecting and accurately locating objects of interest in video sequences. With the rapid development of deep learning technology, visual object tracking has made significant progress in feature modelling, temporal association, and end-to-end learning. This paper systematically reviews the current research status of deep learning-based visual object tracking, analysing the basic ideas and implementation mechanisms under different frameworks from the perspectives of single object tracking (SOT) and multi-object tracking (MOT), and explaining the evolution from traditional hand-crafted feature methods to deep feature-driven models. The paper also summarises commonly used tracking datasets and performance evaluation metrics, and discusses the main challenges faced by current research, including poor long-term modelling capabilities, weak robustness in complex scene tracking, and limited cross-scene generalisation. Finally, the paper looks forward to the future development trends of visual object tracking, highlighting that a unified deep framework integrating multimodal information and spatiotemporal modelling will become an important research direction in the field.