Global Attention Mechanism: Retain Information to Enhance Channel-Spatial Interactions
Global Attention Mechanism: Retain Information to Enhance Channel-Spatial Interactions
About this item
Full title
Author / Creator
Publisher
Ithaca: Cornell University Library, arXiv.org
Journal title
Language
English
Formats
Publication information
Publisher
Ithaca: Cornell University Library, arXiv.org
Subjects
More information
Scope and Contents
Contents
A variety of attention mechanisms have been studied to improve the performance of various computer vision tasks. However, the prior methods overlooked the significance of retaining the information on both channel and spatial aspects to enhance the cross-dimension interactions. Therefore, we propose a global attention mechanism that boosts the perfo...
Alternative Titles
Full title
Global Attention Mechanism: Retain Information to Enhance Channel-Spatial Interactions
Authors, Artists and Contributors
Author / Creator
Identifiers
Primary Identifiers
Record Identifier
TN_cdi_proquest_journals_2609529130
Permalink
https://devfeature-collection.sl.nsw.gov.au/record/TN_cdi_proquest_journals_2609529130
Other Identifiers
E-ISSN
2331-8422
