Learning Soft Mask Based Feature Fusion with Channel and Spatial Attention for Robust Visual Object Tracking

Learning Soft Mask Based Feature Fusion with Channel and Spatial Attention for Robust Visual Object Tracking

We propose to improve the visual object tracking by introducing a soft mask based low-level feature fusion technique. The proposed technique is further strengthened by integrating channel and spatial attention mechanisms. The proposed approach is integrated within a Siamese framework to demonstrate...

Full description

Bibliographic Details
Main Authors: Mustansar Fiaz, Arif Mahmood, Soon Ki Jung
Format: Article
Language:English
Published: MDPI AG 2020-07-01
Series:Sensors
Subjects:
Siamese networks
convolutional neural network
visual tracking
attentional mechanism
Online Access:https://www.mdpi.com/1424-8220/20/14/4021

Internet

https://www.mdpi.com/1424-8220/20/14/4021

Similar Items