Sparse Transformer-Based Sequence Generation for Visual Object Tracking

In visual object tracking, attention mechanisms can flexibly and efficiently handle complex dependencies and global information, which improves tracking accuracy. However, when dealing with scenarios that contain a large amount of background information or other complex information, its global atten...

Full description

Bibliographic Details

Published in:	IEEE Access
Main Authors:	Dan Tian, Dong-Xin Liu, Xiao Wang, Ying Hao
Format:	Article
Language:	English
Published:	IEEE 2024-01-01
Subjects:	Attention mechanism sequence generation sparse attention visual object tracking vision transformer
Online Access:	https://ieeexplore.ieee.org/document/10720760/