Spatial-spectral classification of hyperspectral images: a deep learning framework with Markov random fields based modeling

Chunmei Qing; Jiawei Ruan; Xiangmin Xu; Jinchang Ren; Jaime Zabalza

doi:10.1049/iet-ipr.2018.5727

Spatial-spectral classification of hyperspectral images

a deep learning framework with Markov random fields based modeling

Chunmei Qing, Jiawei Ruan, Xiangmin Xu, Jinchang Ren, Jaime Zabalza

Electronic And Electrical Engineering

Research output: Contribution to journal › Special issue

7 Citations (Scopus)

12 Downloads (Pure)

Abstract

For spatial-spectral classification of hyperspectral images (HSI), a deep learning framework is proposed in this paper, which consists of convolutional neural networks (CNN) and Markov random fields (MRF). Firstly, a CNN model to learn the deep spectral feature from the HSI is built and the class posterior probability distribution is estimated. The CNN with a dropout layer can relieve the overfitting in classification. The CNN is utilized as a pixel-classifier, so it only works in the spectral domain. Then, the spatial information will be encoded by MRF-based multilevel logistic (MLL) prior for regularizing the classification. To derive the correlation of both spectral and spatial features for improving algorithm performance, the marginal probability distribution in HSI is learned using MRF-based loopy belief propagation (LBP). In comparison with several state-of-the-art approaches for data classification on 3 publicly available HSI datasets, experimental results have demonstrated the superior performance of the proposed methodology.

Original language	English
Pages (from-to)	235-245
Number of pages	11
Journal	IET Image Processing
Volume	13
Issue number	2
Early online date	11 Sep 2018
DOIs	https://doi.org/10.1049/iet-ipr.2018.5727
Publication status	Published - 25 Feb 2019

Fingerprint

Neural networks

Probability distributions

Logistics

Classifiers

Pixels

Deep learning

Keywords

hyperspectral image
spatial-spectral classification
convolutional neural networks
Markov random fields
loopy belief propagation

Cite this

Qing, C., Ruan, J., Xu, X., Ren, J., & Zabalza, J. (2019). Spatial-spectral classification of hyperspectral images: a deep learning framework with Markov random fields based modeling. IET Image Processing, 13(2), 235-245. https://doi.org/10.1049/iet-ipr.2018.5727

Qing, Chunmei ; Ruan, Jiawei ; Xu, Xiangmin ; Ren, Jinchang ; Zabalza, Jaime. / Spatial-spectral classification of hyperspectral images : a deep learning framework with Markov random fields based modeling. In: IET Image Processing. 2019 ; Vol. 13, No. 2. pp. 235-245.

@article{1bd85d7108974201b6d5a3b2b151178a,

title = "Spatial-spectral classification of hyperspectral images: a deep learning framework with Markov random fields based modeling",

abstract = "For spatial-spectral classification of hyperspectral images (HSI), a deep learning framework is proposed in this paper, which consists of convolutional neural networks (CNN) and Markov random fields (MRF). Firstly, a CNN model to learn the deep spectral feature from the HSI is built and the class posterior probability distribution is estimated. The CNN with a dropout layer can relieve the overfitting in classification. The CNN is utilized as a pixel-classifier, so it only works in the spectral domain. Then, the spatial information will be encoded by MRF-based multilevel logistic (MLL) prior for regularizing the classification. To derive the correlation of both spectral and spatial features for improving algorithm performance, the marginal probability distribution in HSI is learned using MRF-based loopy belief propagation (LBP). In comparison with several state-of-the-art approaches for data classification on 3 publicly available HSI datasets, experimental results have demonstrated the superior performance of the proposed methodology.",

keywords = "hyperspectral image, spatial-spectral classification, convolutional neural networks, Markov random fields, loopy belief propagation",

author = "Chunmei Qing and Jiawei Ruan and Xiangmin Xu and Jinchang Ren and Jaime Zabalza",

note = "This paper is a postprint of a paper submitted to and accepted for publication in IET Image Processing and is subject to Institution of Engineering and Technology Copyright. The copy of record is available at IET Digital Library.",

year = "2019",

month = "2",

day = "25",

doi = "10.1049/iet-ipr.2018.5727",

language = "English",

volume = "13",

pages = "235--245",

journal = "IET Image Processing",

issn = "1751-9659",

number = "2",

}

Qing, C, Ruan, J, Xu, X, Ren, J & Zabalza, J 2019, 'Spatial-spectral classification of hyperspectral images: a deep learning framework with Markov random fields based modeling', IET Image Processing, vol. 13, no. 2, pp. 235-245. https://doi.org/10.1049/iet-ipr.2018.5727

Spatial-spectral classification of hyperspectral images : a deep learning framework with Markov random fields based modeling. / Qing, Chunmei; Ruan, Jiawei; Xu, Xiangmin; Ren, Jinchang ; Zabalza, Jaime.

In: IET Image Processing, Vol. 13, No. 2, 25.02.2019, p. 235-245.

Research output: Contribution to journal › Special issue

TY - JOUR

T1 - Spatial-spectral classification of hyperspectral images

T2 - a deep learning framework with Markov random fields based modeling

AU - Qing, Chunmei

AU - Ruan, Jiawei

AU - Xu, Xiangmin

AU - Ren, Jinchang

AU - Zabalza, Jaime

N1 - This paper is a postprint of a paper submitted to and accepted for publication in IET Image Processing and is subject to Institution of Engineering and Technology Copyright. The copy of record is available at IET Digital Library.

PY - 2019/2/25

Y1 - 2019/2/25

N2 - For spatial-spectral classification of hyperspectral images (HSI), a deep learning framework is proposed in this paper, which consists of convolutional neural networks (CNN) and Markov random fields (MRF). Firstly, a CNN model to learn the deep spectral feature from the HSI is built and the class posterior probability distribution is estimated. The CNN with a dropout layer can relieve the overfitting in classification. The CNN is utilized as a pixel-classifier, so it only works in the spectral domain. Then, the spatial information will be encoded by MRF-based multilevel logistic (MLL) prior for regularizing the classification. To derive the correlation of both spectral and spatial features for improving algorithm performance, the marginal probability distribution in HSI is learned using MRF-based loopy belief propagation (LBP). In comparison with several state-of-the-art approaches for data classification on 3 publicly available HSI datasets, experimental results have demonstrated the superior performance of the proposed methodology.

AB - For spatial-spectral classification of hyperspectral images (HSI), a deep learning framework is proposed in this paper, which consists of convolutional neural networks (CNN) and Markov random fields (MRF). Firstly, a CNN model to learn the deep spectral feature from the HSI is built and the class posterior probability distribution is estimated. The CNN with a dropout layer can relieve the overfitting in classification. The CNN is utilized as a pixel-classifier, so it only works in the spectral domain. Then, the spatial information will be encoded by MRF-based multilevel logistic (MLL) prior for regularizing the classification. To derive the correlation of both spectral and spatial features for improving algorithm performance, the marginal probability distribution in HSI is learned using MRF-based loopy belief propagation (LBP). In comparison with several state-of-the-art approaches for data classification on 3 publicly available HSI datasets, experimental results have demonstrated the superior performance of the proposed methodology.

KW - hyperspectral image

KW - spatial-spectral classification

KW - convolutional neural networks

KW - Markov random fields

KW - loopy belief propagation

UR - http://digital-library.theiet.org/content/journals/iet-ipr

U2 - 10.1049/iet-ipr.2018.5727

DO - 10.1049/iet-ipr.2018.5727

M3 - Special issue

VL - 13

SP - 235

EP - 245