Recurrent neural networks for remote sensing image classification

Mohamed Ilyes Lakhal, Hakan Çevikalp, Sergio Escalera, Ferda Ofli

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Automatically classifying an image has been a central problem in computer vision for decades. A plethora of models has been proposed, from handcrafted feature solutions to more sophisticated approaches such as deep learning. The authors address the problem of remote sensing image classification, which is an important problem to many real world applications. They introduce a novel deep recurrent architecture that incorporates high-level feature descriptors to tackle this challenging problem. Their solution is based on the general encoder-decoder framework. To the best of the authors' knowledge, this is the first study to use a recurrent network structure on this task. The experimental results show that the proposed framework outperforms the previous works in the three datasets widely used in the literature. They have achieved a state-of-the-art accuracy rate of 97.29% on the UC Merced dataset.

Original languageEnglish
Pages (from-to)1040-1045
Number of pages6
JournalIET Computer Vision
Volume12
Issue number7
DOIs
Publication statusPublished - 1 Oct 2018

Fingerprint

Recurrent neural networks
Image classification
Remote sensing
Computer vision
Deep learning

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this

Recurrent neural networks for remote sensing image classification. / Lakhal, Mohamed Ilyes; Çevikalp, Hakan; Escalera, Sergio; Ofli, Ferda.

In: IET Computer Vision, Vol. 12, No. 7, 01.10.2018, p. 1040-1045.

Research output: Contribution to journalArticle

Lakhal, Mohamed Ilyes ; Çevikalp, Hakan ; Escalera, Sergio ; Ofli, Ferda. / Recurrent neural networks for remote sensing image classification. In: IET Computer Vision. 2018 ; Vol. 12, No. 7. pp. 1040-1045.
@article{b39f336e345b4eac861dd87313b6ae60,
title = "Recurrent neural networks for remote sensing image classification",
abstract = "Automatically classifying an image has been a central problem in computer vision for decades. A plethora of models has been proposed, from handcrafted feature solutions to more sophisticated approaches such as deep learning. The authors address the problem of remote sensing image classification, which is an important problem to many real world applications. They introduce a novel deep recurrent architecture that incorporates high-level feature descriptors to tackle this challenging problem. Their solution is based on the general encoder-decoder framework. To the best of the authors' knowledge, this is the first study to use a recurrent network structure on this task. The experimental results show that the proposed framework outperforms the previous works in the three datasets widely used in the literature. They have achieved a state-of-the-art accuracy rate of 97.29{\%} on the UC Merced dataset.",
author = "Lakhal, {Mohamed Ilyes} and Hakan {\cC}evikalp and Sergio Escalera and Ferda Ofli",
year = "2018",
month = "10",
day = "1",
doi = "10.1049/iet-cvi.2017.0420",
language = "English",
volume = "12",
pages = "1040--1045",
journal = "IET Computer Vision",
issn = "1751-9632",
publisher = "Institution of Engineering and Technology",
number = "7",

}

TY - JOUR

T1 - Recurrent neural networks for remote sensing image classification

AU - Lakhal, Mohamed Ilyes

AU - Çevikalp, Hakan

AU - Escalera, Sergio

AU - Ofli, Ferda

PY - 2018/10/1

Y1 - 2018/10/1

N2 - Automatically classifying an image has been a central problem in computer vision for decades. A plethora of models has been proposed, from handcrafted feature solutions to more sophisticated approaches such as deep learning. The authors address the problem of remote sensing image classification, which is an important problem to many real world applications. They introduce a novel deep recurrent architecture that incorporates high-level feature descriptors to tackle this challenging problem. Their solution is based on the general encoder-decoder framework. To the best of the authors' knowledge, this is the first study to use a recurrent network structure on this task. The experimental results show that the proposed framework outperforms the previous works in the three datasets widely used in the literature. They have achieved a state-of-the-art accuracy rate of 97.29% on the UC Merced dataset.

AB - Automatically classifying an image has been a central problem in computer vision for decades. A plethora of models has been proposed, from handcrafted feature solutions to more sophisticated approaches such as deep learning. The authors address the problem of remote sensing image classification, which is an important problem to many real world applications. They introduce a novel deep recurrent architecture that incorporates high-level feature descriptors to tackle this challenging problem. Their solution is based on the general encoder-decoder framework. To the best of the authors' knowledge, this is the first study to use a recurrent network structure on this task. The experimental results show that the proposed framework outperforms the previous works in the three datasets widely used in the literature. They have achieved a state-of-the-art accuracy rate of 97.29% on the UC Merced dataset.

UR - http://www.scopus.com/inward/record.url?scp=85053455457&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053455457&partnerID=8YFLogxK

U2 - 10.1049/iet-cvi.2017.0420

DO - 10.1049/iet-cvi.2017.0420

M3 - Article

VL - 12

SP - 1040

EP - 1045

JO - IET Computer Vision

JF - IET Computer Vision

SN - 1751-9632

IS - 7

ER -