Remote Sensing Image Scene Classification Based on Global Self-Attention Module

The complexity of scene images makes the research on remote-sensing image scene classification challenging. With the wide application of deep learning in recent years, many remote-sensing scene classification methods using a convolutional neural network (CNN) have emerged. Current CNN usually output...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Qingwen Li, Dongmei Yan, Wanrong Wu
Formato: article
Lenguaje:EN
Publicado: MDPI AG 2021
Materias:
Q
Acceso en línea:https://doaj.org/article/4d717f820e474328b343a656b62df52c
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:4d717f820e474328b343a656b62df52c
record_format dspace
spelling oai:doaj.org-article:4d717f820e474328b343a656b62df52c2021-11-25T18:54:06ZRemote Sensing Image Scene Classification Based on Global Self-Attention Module10.3390/rs132245422072-4292https://doaj.org/article/4d717f820e474328b343a656b62df52c2021-11-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/22/4542https://doaj.org/toc/2072-4292The complexity of scene images makes the research on remote-sensing image scene classification challenging. With the wide application of deep learning in recent years, many remote-sensing scene classification methods using a convolutional neural network (CNN) have emerged. Current CNN usually output global information by integrating the depth features extricated from the convolutional layer through the fully connected layer; however, the global information extracted is not comprehensive. This paper proposes an improved remote-sensing image scene classification method based on a global self-attention module to address this problem. The global information is derived from the depth characteristics extracted by the CNN. In order to better express the semantic information of the remote-sensing image, the multi-head self-attention module is introduced for global information augmentation. Meanwhile, the local perception unit is utilized to improve the self-attention module’s representation capabilities for local objects. The proposed method’s effectiveness is validated through comparative experiments with various training ratios and different scales on public datasets (UC Merced, AID, and NWPU-NESISC45). The precision of our proposed model is significantly improved compared to other methods for remote-sensing image scene classification.Qingwen LiDongmei YanWanrong WuMDPI AGarticleremote-sensing imagescene classificationconvolutional neural network (CNN)global self-attention moduleScienceQENRemote Sensing, Vol 13, Iss 4542, p 4542 (2021)
institution DOAJ
collection DOAJ
language EN
topic remote-sensing image
scene classification
convolutional neural network (CNN)
global self-attention module
Science
Q
spellingShingle remote-sensing image
scene classification
convolutional neural network (CNN)
global self-attention module
Science
Q
Qingwen Li
Dongmei Yan
Wanrong Wu
Remote Sensing Image Scene Classification Based on Global Self-Attention Module
description The complexity of scene images makes the research on remote-sensing image scene classification challenging. With the wide application of deep learning in recent years, many remote-sensing scene classification methods using a convolutional neural network (CNN) have emerged. Current CNN usually output global information by integrating the depth features extricated from the convolutional layer through the fully connected layer; however, the global information extracted is not comprehensive. This paper proposes an improved remote-sensing image scene classification method based on a global self-attention module to address this problem. The global information is derived from the depth characteristics extracted by the CNN. In order to better express the semantic information of the remote-sensing image, the multi-head self-attention module is introduced for global information augmentation. Meanwhile, the local perception unit is utilized to improve the self-attention module’s representation capabilities for local objects. The proposed method’s effectiveness is validated through comparative experiments with various training ratios and different scales on public datasets (UC Merced, AID, and NWPU-NESISC45). The precision of our proposed model is significantly improved compared to other methods for remote-sensing image scene classification.
format article
author Qingwen Li
Dongmei Yan
Wanrong Wu
author_facet Qingwen Li
Dongmei Yan
Wanrong Wu
author_sort Qingwen Li
title Remote Sensing Image Scene Classification Based on Global Self-Attention Module
title_short Remote Sensing Image Scene Classification Based on Global Self-Attention Module
title_full Remote Sensing Image Scene Classification Based on Global Self-Attention Module
title_fullStr Remote Sensing Image Scene Classification Based on Global Self-Attention Module
title_full_unstemmed Remote Sensing Image Scene Classification Based on Global Self-Attention Module
title_sort remote sensing image scene classification based on global self-attention module
publisher MDPI AG
publishDate 2021
url https://doaj.org/article/4d717f820e474328b343a656b62df52c
work_keys_str_mv AT qingwenli remotesensingimagesceneclassificationbasedonglobalselfattentionmodule
AT dongmeiyan remotesensingimagesceneclassificationbasedonglobalselfattentionmodule
AT wanrongwu remotesensingimagesceneclassificationbasedonglobalselfattentionmodule
_version_ 1718410584999854080