GPCA-SIFT: A new local feature descriptor for scene image classification

Lei Ju; Ke Xie; Hao Zheng; Baochang Zhang; Wankou Yang

doi:10.1007/978-981-10-3005-5_24

GPCA-SIFT: A new local feature descriptor for scene image classification

Lei Ju
, Ke Xie
, Hao Zheng
, Baochang Zhang
, Wankou Yang^*

^*Corresponding author for this work

School of Artificial Intelligence(Institute of Artificial Intelligence)

Southeast University, Nanjing
Nanjing Xiaozhuang College

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

3 Scopus citations

Abstract

In this paper, a new local feature descriptor called GPCASIFT is proposed for scene image classification. Like PCA-SIFT, we get the key points using the detection method in Scale Invariant Feature Transform (SIFT) and extract a 41 * 41 patch for each key point. Then we calculate the horizontal and vertical gradient of each pixel in the patch. However, instead of concatenating two gradient matrices, we directly work with the two-dimensional matrix and apply Generalized Principal Component Analysis (GPCA) to reduce it to a lower-dimensional matrix. Finally, we concatenate the reduced matrix and form a 1D vector. Compared with Principal Component Analysis (PCA), it preserves more spatial locality information. When applied in multi-class scene image classification, our proposed descriptor outperforms other related algorithms in terms of classification accuracy.

Original language	English
Title of host publication	Pattern Recognition - 7th Chinese Conference, CCPR 2016, Proceedings
Editors	Tieniu Tan, Xilin Chen, Xuelong Li, Jian Yang, Hong Cheng, Jie Zhou
Publisher	Springer Verlag
Pages	286-295
Number of pages	10
ISBN (Print)	9789811030048
DOIs	https://doi.org/10.1007/978-981-10-3005-5_24
State	Published - 2016

Publication series

Name	Communications in Computer and Information Science
Volume	663
ISSN (Print)	1865-0929

Keywords

GPCA
GPCA-SIFT
SIFT
Scene image classification

Access to Document

10.1007/978-981-10-3005-5_24

Cite this

Ju, L., Xie, K., Zheng, H., Zhang, B., & Yang, W. (2016). GPCA-SIFT: A new local feature descriptor for scene image classification. In T. Tan, X. Chen, X. Li, J. Yang, H. Cheng, & J. Zhou (Eds.), Pattern Recognition - 7th Chinese Conference, CCPR 2016, Proceedings (pp. 286-295). (Communications in Computer and Information Science; Vol. 663). Springer Verlag. https://doi.org/10.1007/978-981-10-3005-5_24

@inproceedings{8a06818872364f1d80a759174c0165db,

title = "GPCA-SIFT: A new local feature descriptor for scene image classification",

abstract = "In this paper, a new local feature descriptor called GPCASIFT is proposed for scene image classification. Like PCA-SIFT, we get the key points using the detection method in Scale Invariant Feature Transform (SIFT) and extract a 41 * 41 patch for each key point. Then we calculate the horizontal and vertical gradient of each pixel in the patch. However, instead of concatenating two gradient matrices, we directly work with the two-dimensional matrix and apply Generalized Principal Component Analysis (GPCA) to reduce it to a lower-dimensional matrix. Finally, we concatenate the reduced matrix and form a 1D vector. Compared with Principal Component Analysis (PCA), it preserves more spatial locality information. When applied in multi-class scene image classification, our proposed descriptor outperforms other related algorithms in terms of classification accuracy.",

keywords = "GPCA, GPCA-SIFT, SIFT, Scene image classification",

author = "Lei Ju and Ke Xie and Hao Zheng and Baochang Zhang and Wankou Yang",

note = "Publisher Copyright: {\textcopyright} Springer Nature Singapore Pte Ltd. 2016.",

year = "2016",

doi = "10.1007/978-981-10-3005-5\_24",

language = "英语",

isbn = "9789811030048",

series = "Communications in Computer and Information Science",

publisher = "Springer Verlag",

pages = "286--295",

editor = "Tieniu Tan and Xilin Chen and Xuelong Li and Jian Yang and Hong Cheng and Jie Zhou",

booktitle = "Pattern Recognition - 7th Chinese Conference, CCPR 2016, Proceedings",

address = "德国",

}

Ju, L, Xie, K, Zheng, H, Zhang, B & Yang, W 2016, GPCA-SIFT: A new local feature descriptor for scene image classification. in T Tan, X Chen, X Li, J Yang, H Cheng & J Zhou (eds), Pattern Recognition - 7th Chinese Conference, CCPR 2016, Proceedings. Communications in Computer and Information Science, vol. 663, Springer Verlag, pp. 286-295. https://doi.org/10.1007/978-981-10-3005-5_24

GPCA-SIFT: A new local feature descriptor for scene image classification. / Ju, Lei; Xie, Ke; Zheng, Hao et al.
Pattern Recognition - 7th Chinese Conference, CCPR 2016, Proceedings. ed. / Tieniu Tan; Xilin Chen; Xuelong Li; Jian Yang; Hong Cheng; Jie Zhou. Springer Verlag, 2016. p. 286-295 (Communications in Computer and Information Science; Vol. 663).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - GPCA-SIFT

T2 - A new local feature descriptor for scene image classification

AU - Ju, Lei

AU - Xie, Ke

AU - Zheng, Hao

AU - Zhang, Baochang

AU - Yang, Wankou

N1 - Publisher Copyright: © Springer Nature Singapore Pte Ltd. 2016.

PY - 2016

Y1 - 2016

N2 - In this paper, a new local feature descriptor called GPCASIFT is proposed for scene image classification. Like PCA-SIFT, we get the key points using the detection method in Scale Invariant Feature Transform (SIFT) and extract a 41 * 41 patch for each key point. Then we calculate the horizontal and vertical gradient of each pixel in the patch. However, instead of concatenating two gradient matrices, we directly work with the two-dimensional matrix and apply Generalized Principal Component Analysis (GPCA) to reduce it to a lower-dimensional matrix. Finally, we concatenate the reduced matrix and form a 1D vector. Compared with Principal Component Analysis (PCA), it preserves more spatial locality information. When applied in multi-class scene image classification, our proposed descriptor outperforms other related algorithms in terms of classification accuracy.

AB - In this paper, a new local feature descriptor called GPCASIFT is proposed for scene image classification. Like PCA-SIFT, we get the key points using the detection method in Scale Invariant Feature Transform (SIFT) and extract a 41 * 41 patch for each key point. Then we calculate the horizontal and vertical gradient of each pixel in the patch. However, instead of concatenating two gradient matrices, we directly work with the two-dimensional matrix and apply Generalized Principal Component Analysis (GPCA) to reduce it to a lower-dimensional matrix. Finally, we concatenate the reduced matrix and form a 1D vector. Compared with Principal Component Analysis (PCA), it preserves more spatial locality information. When applied in multi-class scene image classification, our proposed descriptor outperforms other related algorithms in terms of classification accuracy.

KW - GPCA

KW - GPCA-SIFT

KW - SIFT

KW - Scene image classification

UR - https://www.scopus.com/pages/publications/84994888499

U2 - 10.1007/978-981-10-3005-5_24

DO - 10.1007/978-981-10-3005-5_24

M3 - 会议稿件

AN - SCOPUS:84994888499

SN - 9789811030048

T3 - Communications in Computer and Information Science

SP - 286

EP - 295

BT - Pattern Recognition - 7th Chinese Conference, CCPR 2016, Proceedings

A2 - Tan, Tieniu

A2 - Chen, Xilin

A2 - Li, Xuelong

A2 - Yang, Jian

A2 - Cheng, Hong

A2 - Zhou, Jie

PB - Springer Verlag

ER -

GPCA-SIFT: A new local feature descriptor for scene image classification

Abstract

Publication series

Keywords

Access to Document

Other files and links

Fingerprint

Cite this