The tracking methods based on Transformer have shown great potential in visual tracking
and achieved significant tracking performance. The traditional transformer based feature …

A memory mechanism has attracted growing popularity in tracking tasks due to the ability of
learning long-term-dependent information. However, it is very challenging for existing …

Transformer-based trackers greatly improve tracking success rate and precision rate.
Attention mechanism in Transformer can fully explore the context information across …

The traditional Siamese network based object tracking algorithms suffer from high
computational complexity, making them difficult to run on embedded devices. Moreover …