A New Convolutional Kernel Classifier for Hyperspectral Image Classification

Multiplekernel learning (MKL) algorithms are among the most successful classification methods for hyperspectral data. Nevertheless, these algorithms suffer from two main drawbacks of computational complexity and debility to admit to the end-to-end learning paradigm. This article proposed a convoluti...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Mohsen Ansari, Saeid Homayouni, Abdolreza Safari, Saeid Niazmardi
Formato: article
Lenguaje:EN
Publicado: IEEE 2021
Materias:
Acceso en línea:https://doaj.org/article/fd4f579c698548bc9486964d61cb91d0
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:fd4f579c698548bc9486964d61cb91d0
record_format dspace
spelling oai:doaj.org-article:fd4f579c698548bc9486964d61cb91d02021-11-17T00:00:16ZA New Convolutional Kernel Classifier for Hyperspectral Image Classification2151-153510.1109/JSTARS.2021.3123087https://doaj.org/article/fd4f579c698548bc9486964d61cb91d02021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9591367/https://doaj.org/toc/2151-1535Multiplekernel learning (MKL) algorithms are among the most successful classification methods for hyperspectral data. Nevertheless, these algorithms suffer from two main drawbacks of computational complexity and debility to admit to the end-to-end learning paradigm. This article proposed a convolutional kernel classifier (CKC) for hyperspectral remote sensing images to address these issues. The CKC uses the Nystr&#x00F6;m approximation method to estimate a low-rank approximation of the basis kernels, thus solves the issues associated with the high dimensionality of the basis kernels. The CKC uses deep architecture to learn the optimal combination of the basis kernels and the classification task to enable end-to-end learning. The proposed CKC&#x0027;s architecture is based on a one-dimensional-convolutional neural network (CNN-1-D), and it uses kernel dropout to prevent overfitting. It is the first instance of deep-kernel algorithms in the field of remote sensing. The proposed method was compared with several well-known hyperspectral image analysis MKL algorithms, including a multi-kernel variant of the deep kernel machine optimization, MKL-average, Simple-MKL, and generalize MKL, and state-of-the-art deep learning models, including Vanilla recurrent neural network (VanillaRNN) and CNN-1-D in classifying four benchmark hyperspectral datasets. The experimental results show that the CKC consistently outperforms all the competitor methods, and its runtime is lower than its MKL algorithm counterparts on four benchmark hyperspectral datasets. Moreover, the Nystr&#x00F6;m approximation solves the high dimensionality of the basis kernels and boosts classification accuracy. The source codes of CKC are available from: <uri>https://github.com/MohsenAnsari1373/A-New-Convolutional-Kernel-Classifier-for-Hyperspectral-Image-Classification</uri>.Mohsen AnsariSaeid HomayouniAbdolreza SafariSaeid NiazmardiIEEEarticleConvolutional neural network (CNN)deep kernelhyperspectral classificationmultiple kernel learning (MKL)Ocean engineeringTC1501-1800Geophysics. Cosmic physicsQC801-809ENIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol 14, Pp 11240-11256 (2021)
institution DOAJ
collection DOAJ
language EN
topic Convolutional neural network (CNN)
deep kernel
hyperspectral classification
multiple kernel learning (MKL)
Ocean engineering
TC1501-1800
Geophysics. Cosmic physics
QC801-809
spellingShingle Convolutional neural network (CNN)
deep kernel
hyperspectral classification
multiple kernel learning (MKL)
Ocean engineering
TC1501-1800
Geophysics. Cosmic physics
QC801-809
Mohsen Ansari
Saeid Homayouni
Abdolreza Safari
Saeid Niazmardi
A New Convolutional Kernel Classifier for Hyperspectral Image Classification
description Multiplekernel learning (MKL) algorithms are among the most successful classification methods for hyperspectral data. Nevertheless, these algorithms suffer from two main drawbacks of computational complexity and debility to admit to the end-to-end learning paradigm. This article proposed a convolutional kernel classifier (CKC) for hyperspectral remote sensing images to address these issues. The CKC uses the Nystr&#x00F6;m approximation method to estimate a low-rank approximation of the basis kernels, thus solves the issues associated with the high dimensionality of the basis kernels. The CKC uses deep architecture to learn the optimal combination of the basis kernels and the classification task to enable end-to-end learning. The proposed CKC&#x0027;s architecture is based on a one-dimensional-convolutional neural network (CNN-1-D), and it uses kernel dropout to prevent overfitting. It is the first instance of deep-kernel algorithms in the field of remote sensing. The proposed method was compared with several well-known hyperspectral image analysis MKL algorithms, including a multi-kernel variant of the deep kernel machine optimization, MKL-average, Simple-MKL, and generalize MKL, and state-of-the-art deep learning models, including Vanilla recurrent neural network (VanillaRNN) and CNN-1-D in classifying four benchmark hyperspectral datasets. The experimental results show that the CKC consistently outperforms all the competitor methods, and its runtime is lower than its MKL algorithm counterparts on four benchmark hyperspectral datasets. Moreover, the Nystr&#x00F6;m approximation solves the high dimensionality of the basis kernels and boosts classification accuracy. The source codes of CKC are available from: <uri>https://github.com/MohsenAnsari1373/A-New-Convolutional-Kernel-Classifier-for-Hyperspectral-Image-Classification</uri>.
format article
author Mohsen Ansari
Saeid Homayouni
Abdolreza Safari
Saeid Niazmardi
author_facet Mohsen Ansari
Saeid Homayouni
Abdolreza Safari
Saeid Niazmardi
author_sort Mohsen Ansari
title A New Convolutional Kernel Classifier for Hyperspectral Image Classification
title_short A New Convolutional Kernel Classifier for Hyperspectral Image Classification
title_full A New Convolutional Kernel Classifier for Hyperspectral Image Classification
title_fullStr A New Convolutional Kernel Classifier for Hyperspectral Image Classification
title_full_unstemmed A New Convolutional Kernel Classifier for Hyperspectral Image Classification
title_sort new convolutional kernel classifier for hyperspectral image classification
publisher IEEE
publishDate 2021
url https://doaj.org/article/fd4f579c698548bc9486964d61cb91d0
work_keys_str_mv AT mohsenansari anewconvolutionalkernelclassifierforhyperspectralimageclassification
AT saeidhomayouni anewconvolutionalkernelclassifierforhyperspectralimageclassification
AT abdolrezasafari anewconvolutionalkernelclassifierforhyperspectralimageclassification
AT saeidniazmardi anewconvolutionalkernelclassifierforhyperspectralimageclassification
AT mohsenansari newconvolutionalkernelclassifierforhyperspectralimageclassification
AT saeidhomayouni newconvolutionalkernelclassifierforhyperspectralimageclassification
AT abdolrezasafari newconvolutionalkernelclassifierforhyperspectralimageclassification
AT saeidniazmardi newconvolutionalkernelclassifierforhyperspectralimageclassification
_version_ 1718426086381977600