.png){kind=icon}
返回
Transformer tracking with multi-scale extended attention
.png){kind=icon}
PRE
AI
DOI：10.1016/j.image.2026.117503.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
• A Multi-Scale Extended Attention (MSEA) based backbone network is designed. • MSEA captures multi-scale information with lower computational redundancy. • A Transformer backbone network is designed with MSEA and Multi-Head Self-attention. • A dual-branch dual-stage tracker is proposed based on the backbone network.
Keyword:
Multi-Scale Extended Attention
Transformer Backbone
Dual-Branch Tracker
Multi-Head Self-Attention
Object Tracking
期刊
S
IF:
0
论文数:
71
被引数:
0
机构
暂无机构信息
引用论文
.png){kind=icon}
分享收藏Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition用于视觉识别的深度卷积网络中的空间金字塔池化
OAAI
OAAI分享收藏
查看更多内容