Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation

Shenmiao Han, Zhensen Wu, Guangli Wu, Jun Tan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

In this paper, we propose an aurora classification method using a multi-level feature representation aimed to capture both global and local texture information, and to reduce the feature space dimension substantially. First-order and second-order statistics are computed for an input image and its low-frequency scaled images at three lower levels obtained using wavelet decomposition. The features include gray level distribution, co-occurrence matrix features, and run-length matrix features. A support vector machine (SVM) classifier was trained and tested on a Chinese Arctic Yellow River Station dayside aurora image dataset. Classification performance was evaluated and compared with those of k-nearest neighbor (KNN) classifiers and back-propagation neural networks (BPNN). To explore the possibility of using a smaller feature space, we used a Minimum-Redundancy Max-Relevance feature selection strategy. The result shows that there is only indistinct performance decrease by reducing the feature vector from a total of 88 to the most discriminatory 38 features. This proves that our multi-level feature representation is very robust.

Original languageEnglish (US)
Title of host publicationMaterial and Manufacturing Technology II
Pages158-162
Number of pages5
Volume341-342
DOIs
StatePublished - Jan 1 2012
Event2011 2nd International Conference on Material and Manufacturing Technology, ICMMT 2011 - Xiamen, China
Duration: Jul 8 2011Jul 11 2011

Publication series

NameAdvanced Materials Research
Volume341-342
ISSN (Print)1022-6680

Other

Other2011 2nd International Conference on Material and Manufacturing Technology, ICMMT 2011
CountryChina
CityXiamen
Period7/8/117/11/11

Fingerprint

Classifiers
Textures
Wavelet decomposition
Backpropagation
Support vector machines
Redundancy
Feature extraction
Rivers
Statistics
Neural networks

Keywords

  • Aurora
  • Gray level co-occurrence matrix
  • Run-length
  • Support vector machine
  • Wavelet decomposition

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Han, S., Wu, Z., Wu, G., & Tan, J. (2012). Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation. In Material and Manufacturing Technology II (Vol. 341-342, pp. 158-162). (Advanced Materials Research; Vol. 341-342). https://doi.org/10.4028/www.scientific.net/AMR.341-342.158

Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation. / Han, Shenmiao; Wu, Zhensen; Wu, Guangli; Tan, Jun.

Material and Manufacturing Technology II. Vol. 341-342 2012. p. 158-162 (Advanced Materials Research; Vol. 341-342).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Han, S, Wu, Z, Wu, G & Tan, J 2012, Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation. in Material and Manufacturing Technology II. vol. 341-342, Advanced Materials Research, vol. 341-342, pp. 158-162, 2011 2nd International Conference on Material and Manufacturing Technology, ICMMT 2011, Xiamen, China, 7/8/11. https://doi.org/10.4028/www.scientific.net/AMR.341-342.158
Han S, Wu Z, Wu G, Tan J. Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation. In Material and Manufacturing Technology II. Vol. 341-342. 2012. p. 158-162. (Advanced Materials Research). https://doi.org/10.4028/www.scientific.net/AMR.341-342.158
Han, Shenmiao ; Wu, Zhensen ; Wu, Guangli ; Tan, Jun. / Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation. Material and Manufacturing Technology II. Vol. 341-342 2012. pp. 158-162 (Advanced Materials Research).
@inproceedings{156b4e4db1d245f9b8e603bc7f548aad,
title = "Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation",
abstract = "In this paper, we propose an aurora classification method using a multi-level feature representation aimed to capture both global and local texture information, and to reduce the feature space dimension substantially. First-order and second-order statistics are computed for an input image and its low-frequency scaled images at three lower levels obtained using wavelet decomposition. The features include gray level distribution, co-occurrence matrix features, and run-length matrix features. A support vector machine (SVM) classifier was trained and tested on a Chinese Arctic Yellow River Station dayside aurora image dataset. Classification performance was evaluated and compared with those of k-nearest neighbor (KNN) classifiers and back-propagation neural networks (BPNN). To explore the possibility of using a smaller feature space, we used a Minimum-Redundancy Max-Relevance feature selection strategy. The result shows that there is only indistinct performance decrease by reducing the feature vector from a total of 88 to the most discriminatory 38 features. This proves that our multi-level feature representation is very robust.",
keywords = "Aurora, Gray level co-occurrence matrix, Run-length, Support vector machine, Wavelet decomposition",
author = "Shenmiao Han and Zhensen Wu and Guangli Wu and Jun Tan",
year = "2012",
month = "1",
day = "1",
doi = "10.4028/www.scientific.net/AMR.341-342.158",
language = "English (US)",
isbn = "9783037852521",
volume = "341-342",
series = "Advanced Materials Research",
pages = "158--162",
booktitle = "Material and Manufacturing Technology II",

}

TY - GEN

T1 - Automatic classification of dayside aurora in all-sky images using a multi-level texture feature representation

AU - Han, Shenmiao

AU - Wu, Zhensen

AU - Wu, Guangli

AU - Tan, Jun

PY - 2012/1/1

Y1 - 2012/1/1

N2 - In this paper, we propose an aurora classification method using a multi-level feature representation aimed to capture both global and local texture information, and to reduce the feature space dimension substantially. First-order and second-order statistics are computed for an input image and its low-frequency scaled images at three lower levels obtained using wavelet decomposition. The features include gray level distribution, co-occurrence matrix features, and run-length matrix features. A support vector machine (SVM) classifier was trained and tested on a Chinese Arctic Yellow River Station dayside aurora image dataset. Classification performance was evaluated and compared with those of k-nearest neighbor (KNN) classifiers and back-propagation neural networks (BPNN). To explore the possibility of using a smaller feature space, we used a Minimum-Redundancy Max-Relevance feature selection strategy. The result shows that there is only indistinct performance decrease by reducing the feature vector from a total of 88 to the most discriminatory 38 features. This proves that our multi-level feature representation is very robust.

AB - In this paper, we propose an aurora classification method using a multi-level feature representation aimed to capture both global and local texture information, and to reduce the feature space dimension substantially. First-order and second-order statistics are computed for an input image and its low-frequency scaled images at three lower levels obtained using wavelet decomposition. The features include gray level distribution, co-occurrence matrix features, and run-length matrix features. A support vector machine (SVM) classifier was trained and tested on a Chinese Arctic Yellow River Station dayside aurora image dataset. Classification performance was evaluated and compared with those of k-nearest neighbor (KNN) classifiers and back-propagation neural networks (BPNN). To explore the possibility of using a smaller feature space, we used a Minimum-Redundancy Max-Relevance feature selection strategy. The result shows that there is only indistinct performance decrease by reducing the feature vector from a total of 88 to the most discriminatory 38 features. This proves that our multi-level feature representation is very robust.

KW - Aurora

KW - Gray level co-occurrence matrix

KW - Run-length

KW - Support vector machine

KW - Wavelet decomposition

UR - http://www.scopus.com/inward/record.url?scp=80054005276&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80054005276&partnerID=8YFLogxK

U2 - 10.4028/www.scientific.net/AMR.341-342.158

DO - 10.4028/www.scientific.net/AMR.341-342.158

M3 - Conference contribution

AN - SCOPUS:80054005276

SN - 9783037852521

VL - 341-342

T3 - Advanced Materials Research

SP - 158

EP - 162

BT - Material and Manufacturing Technology II

ER -