Multi-branch convolutional descriptors for content-based remote sensing image retrieval

Raffaele Imbriaco; Tunc Alkanat; Egor Bondarev; Peter H.N. de With

Multi-branch convolutional descriptors for content-based remote sensing image retrieval

Raffaele Imbriaco, Tunc Alkanat, Egor Bondarev, Peter H.N. de With

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

Abstract

Context-based remote sensing image retrieval (CBRSIR) is an important problem in computer vision with many applications such as military, agriculture, and surveillance. In this study, inspired by recent developments in person re-identification, we design and fine-tune a multi-branch deep learning architecture that combines global and local features to obtain rich and discriminative image representations. Additionally, we propose a new evaluation strategy that fully separates the test and training sets and where new unseen data is used for querying, thereby emphasizing the generalization capability of retrieval systems. Extensive evaluations show that our method significantly outperforms the existing approaches by up to 10.7% in mean precision@20 on popular CBRSIR datasets. Regarding the new evaluation strategy, our method attains excellent retrieval performance, yielding more than 95% precision@20 score on the challenging PatternNet dataset.

Original language	English
Title of host publication	VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
Editors	Giovanni Maria Farinella, Petia Radeva, Jose Braz
Publisher	SciTePress Digital Library
Pages	242-249
Number of pages	8
Volume	5
ISBN (Electronic)	9789897584022
Publication status	Published - 1 Jan 2020
Event	15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2020 - Valletta, Malta Duration: 27 Feb 2020 → 29 Feb 2020

Conference

Conference	15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2020
Country/Territory	Malta
City	Valletta
Period	27/02/20 → 29/02/20

Funding

The authors kindly appreciate the NVIDIA gift of a Titan Xp GPU for this research. This work was supported by the Interreg project PASSAnT and the European ITEA project PS-CRIMSON.

Keywords

Content-based Image Retrieval
Convolutional Neural Networks
Local Feature Extraction
Remote Sensing

Cite this

Imbriaco, R., Alkanat, T., Bondarev, E., & de With, P. H. N. (2020). Multi-branch convolutional descriptors for content-based remote sensing image retrieval. In G. M. Farinella, P. Radeva, & J. Braz (Eds.), VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (Vol. 5, pp. 242-249). SciTePress Digital Library.

Imbriaco, Raffaele ; Alkanat, Tunc ; Bondarev, Egor et al. / Multi-branch convolutional descriptors for content-based remote sensing image retrieval. VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. editor / Giovanni Maria Farinella ; Petia Radeva ; Jose Braz. Vol. 5 SciTePress Digital Library, 2020. pp. 242-249

@inproceedings{6cf27d2afa8543edb6daa0ac386ee20d,

title = "Multi-branch convolutional descriptors for content-based remote sensing image retrieval",

abstract = "Context-based remote sensing image retrieval (CBRSIR) is an important problem in computer vision with many applications such as military, agriculture, and surveillance. In this study, inspired by recent developments in person re-identification, we design and fine-tune a multi-branch deep learning architecture that combines global and local features to obtain rich and discriminative image representations. Additionally, we propose a new evaluation strategy that fully separates the test and training sets and where new unseen data is used for querying, thereby emphasizing the generalization capability of retrieval systems. Extensive evaluations show that our method significantly outperforms the existing approaches by up to 10.7\% in mean precision@20 on popular CBRSIR datasets. Regarding the new evaluation strategy, our method attains excellent retrieval performance, yielding more than 95\% precision@20 score on the challenging PatternNet dataset.",

keywords = "Content-based Image Retrieval, Convolutional Neural Networks, Local Feature Extraction, Remote Sensing",

author = "Raffaele Imbriaco and Tunc Alkanat and Egor Bondarev and \{de With\}, \{Peter H.N.\}",

year = "2020",

month = jan,

day = "1",

language = "English",

volume = "5",

pages = "242--249",

editor = "Farinella, \{Giovanni Maria\} and Petia Radeva and Jose Braz",

booktitle = "VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications",

publisher = "SciTePress Digital Library",

note = "15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2020 ; Conference date: 27-02-2020 Through 29-02-2020",

}

Imbriaco, R, Alkanat, T , Bondarev, E & de With, PHN 2020, Multi-branch convolutional descriptors for content-based remote sensing image retrieval. in GM Farinella, P Radeva & J Braz (eds), VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. vol. 5, SciTePress Digital Library, pp. 242-249, 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2020, Valletta, Malta, 27/02/20.

Multi-branch convolutional descriptors for content-based remote sensing image retrieval. / Imbriaco, Raffaele; Alkanat, Tunc ; Bondarev, Egor et al.
VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. ed. / Giovanni Maria Farinella; Petia Radeva; Jose Braz. Vol. 5 SciTePress Digital Library, 2020. p. 242-249.

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

TY - GEN

T1 - Multi-branch convolutional descriptors for content-based remote sensing image retrieval

AU - Imbriaco, Raffaele

AU - Alkanat, Tunc

AU - Bondarev, Egor

AU - de With, Peter H.N.

PY - 2020/1/1

Y1 - 2020/1/1

N2 - Context-based remote sensing image retrieval (CBRSIR) is an important problem in computer vision with many applications such as military, agriculture, and surveillance. In this study, inspired by recent developments in person re-identification, we design and fine-tune a multi-branch deep learning architecture that combines global and local features to obtain rich and discriminative image representations. Additionally, we propose a new evaluation strategy that fully separates the test and training sets and where new unseen data is used for querying, thereby emphasizing the generalization capability of retrieval systems. Extensive evaluations show that our method significantly outperforms the existing approaches by up to 10.7% in mean precision@20 on popular CBRSIR datasets. Regarding the new evaluation strategy, our method attains excellent retrieval performance, yielding more than 95% precision@20 score on the challenging PatternNet dataset.

AB - Context-based remote sensing image retrieval (CBRSIR) is an important problem in computer vision with many applications such as military, agriculture, and surveillance. In this study, inspired by recent developments in person re-identification, we design and fine-tune a multi-branch deep learning architecture that combines global and local features to obtain rich and discriminative image representations. Additionally, we propose a new evaluation strategy that fully separates the test and training sets and where new unseen data is used for querying, thereby emphasizing the generalization capability of retrieval systems. Extensive evaluations show that our method significantly outperforms the existing approaches by up to 10.7% in mean precision@20 on popular CBRSIR datasets. Regarding the new evaluation strategy, our method attains excellent retrieval performance, yielding more than 95% precision@20 score on the challenging PatternNet dataset.

KW - Content-based Image Retrieval

KW - Convolutional Neural Networks

KW - Local Feature Extraction

KW - Remote Sensing

UR - http://www.scopus.com/inward/record.url?scp=85083501381&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85083501381

VL - 5

SP - 242

EP - 249

BT - VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

A2 - Farinella, Giovanni Maria

A2 - Radeva, Petia

A2 - Braz, Jose

PB - SciTePress Digital Library

T2 - 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2020

Y2 - 27 February 2020 through 29 February 2020

ER -

Imbriaco R, Alkanat T , Bondarev E , de With PHN. Multi-branch convolutional descriptors for content-based remote sensing image retrieval. In Farinella GM, Radeva P, Braz J, editors, VISIGRAPP 2020 - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Vol. 5. SciTePress Digital Library. 2020. p. 242-249

Multi-branch convolutional descriptors for content-based remote sensing image retrieval

Abstract

Conference

Funding

Keywords

Other files and links

Fingerprint

Cite this