Deep Learning for Single-object Tracking: A Survey

With the rapid advancement of deep learning, this technique has also shown promise in single-object tracking, considerably enhancing the algorithm's performance. This paper will summarize and analyze the evaluation metrics, public datasets, state-of-the-art tracking methods, etc. It will classify the research status of the key technologies involved and reveal the mystery of single object tracking technology step by step. In addition, the significant challenges and problems of single object tracking technology are summarized.