Knowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification

Due to device limitations, small networks are necessary for some real-world scenarios, such as satellites and micro-robots. Therefore, the development of a network with both good performance and small size is an important area of research. Deep networks can learn well from large amounts of data, whi...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Ling Tian, Zhichao Wang, Bokun He, Chu He, Dingwen Wang, Deshi Li
Formato: article
Lenguaje:EN
Publicado: MDPI AG 2021
Materias:
Q
Acceso en línea:https://doaj.org/article/b502d9b230bd4486a4c0a2f5b86bebe2
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:b502d9b230bd4486a4c0a2f5b86bebe2
record_format dspace
spelling oai:doaj.org-article:b502d9b230bd4486a4c0a2f5b86bebe22021-11-25T18:54:01ZKnowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification10.3390/rs132245372072-4292https://doaj.org/article/b502d9b230bd4486a4c0a2f5b86bebe22021-11-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/22/4537https://doaj.org/toc/2072-4292Due to device limitations, small networks are necessary for some real-world scenarios, such as satellites and micro-robots. Therefore, the development of a network with both good performance and small size is an important area of research. Deep networks can learn well from large amounts of data, while manifold networks have outstanding feature representation at small sizes. In this paper, we propose an approach that exploits the advantages of deep networks and shallow Grassmannian manifold networks. Inspired by knowledge distillation, we use the information learned from convolutional neural networks to guide the training of the manifold networks. Our approach leads to a reduction in model size, which addresses the problem of deploying deep learning on resource-limited embedded devices. Finally, a series of experiments were conducted on four remote sensing scene classification datasets. The method in this paper improved the classification accuracy by 2.31% and 1.73% on the UC Merced Land Use and SIRIWHU datasets, respectively, and the experimental results demonstrate the effectiveness of our approach.Ling TianZhichao WangBokun HeChu HeDingwen WangDeshi LiMDPI AGarticleknowledge distillationGrassmann manifoldneural networkScienceQENRemote Sensing, Vol 13, Iss 4537, p 4537 (2021)
institution DOAJ
collection DOAJ
language EN
topic knowledge distillation
Grassmann manifold
neural network
Science
Q
spellingShingle knowledge distillation
Grassmann manifold
neural network
Science
Q
Ling Tian
Zhichao Wang
Bokun He
Chu He
Dingwen Wang
Deshi Li
Knowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification
description Due to device limitations, small networks are necessary for some real-world scenarios, such as satellites and micro-robots. Therefore, the development of a network with both good performance and small size is an important area of research. Deep networks can learn well from large amounts of data, while manifold networks have outstanding feature representation at small sizes. In this paper, we propose an approach that exploits the advantages of deep networks and shallow Grassmannian manifold networks. Inspired by knowledge distillation, we use the information learned from convolutional neural networks to guide the training of the manifold networks. Our approach leads to a reduction in model size, which addresses the problem of deploying deep learning on resource-limited embedded devices. Finally, a series of experiments were conducted on four remote sensing scene classification datasets. The method in this paper improved the classification accuracy by 2.31% and 1.73% on the UC Merced Land Use and SIRIWHU datasets, respectively, and the experimental results demonstrate the effectiveness of our approach.
format article
author Ling Tian
Zhichao Wang
Bokun He
Chu He
Dingwen Wang
Deshi Li
author_facet Ling Tian
Zhichao Wang
Bokun He
Chu He
Dingwen Wang
Deshi Li
author_sort Ling Tian
title Knowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification
title_short Knowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification
title_full Knowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification
title_fullStr Knowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification
title_full_unstemmed Knowledge Distillation of Grassmann Manifold Network for Remote Sensing Scene Classification
title_sort knowledge distillation of grassmann manifold network for remote sensing scene classification
publisher MDPI AG
publishDate 2021
url https://doaj.org/article/b502d9b230bd4486a4c0a2f5b86bebe2
work_keys_str_mv AT lingtian knowledgedistillationofgrassmannmanifoldnetworkforremotesensingsceneclassification
AT zhichaowang knowledgedistillationofgrassmannmanifoldnetworkforremotesensingsceneclassification
AT bokunhe knowledgedistillationofgrassmannmanifoldnetworkforremotesensingsceneclassification
AT chuhe knowledgedistillationofgrassmannmanifoldnetworkforremotesensingsceneclassification
AT dingwenwang knowledgedistillationofgrassmannmanifoldnetworkforremotesensingsceneclassification
AT deshili knowledgedistillationofgrassmannmanifoldnetworkforremotesensingsceneclassification
_version_ 1718410613941600256