Visual place recognition using landmark distribution descriptors

Pilailuck Panphattarasap; Andrew Calway

doi:10.1007/978-3-319-54190-7_30

Visual place recognition using landmark distribution descriptors

Pilailuck Panphattarasap^*, Andrew Calway

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

12 Citations (Scopus)

293 Downloads (Pure)

Abstract

Recent work by Sünderhauf et al. [1] demonstrated improved visual place recognition using proposal regions coupled with features from convolutional neural networks (CNN) to match landmarks between views. In this work we extend the approach by introducing descriptors built from landmark features which also encode the spatial distribution of the landmarks within a view. Matching descriptors then enforces consistency of the relative positions of landmarks between views. This has a significant impact on performance. For example, in experiments on 10 image-pair datasets, each consisting of 200 urban locations with significant differences in viewing positions and conditions, we recorded average precision of around 70% (at 100% recall), compared with 58% obtained using whole image CNN features and 50% for the method in [1].

Original language	English
Title of host publication	Computer Vision - ACCV 2016
Subtitle of host publication	13th Asian Conference on Computer Vision, ACCV 2016, Revised Selected Papers
Publisher	Springer-Verlag Berlin
Pages	487-502
Number of pages	16
Volume	10114 LNCS
ISBN (Print)	9783319541891
DOIs	https://doi.org/10.1007/978-3-319-54190-7_30
Publication status	Published - 2017
Event	13th Asian Conference on Computer Vision 2016: Workshop on Assistive Vision - Taipei International Convention Center, Taipei, Taiwan Duration: 20 Nov 2016 → 24 Nov 2016 Conference number: 13 http://www.accv2016.org/

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10114 LNCS
ISSN (Print)	03029743
ISSN (Electronic)	16113349

Conference

Conference	13th Asian Conference on Computer Vision 2016
Abbreviated title	ACCV 16
Country/Territory	Taiwan
City	Taipei
Period	20/11/16 → 24/11/16
Internet address	http://www.accv2016.org/

Access to Document

10.1007/978-3-319-54190-7_30

Full-text PDF (accepted author manuscript)
This is the author accepted manuscript (AAM). The final published version (version of record) is available online via SpringerLink at https://link.springer.com/chapter/10.1007%2F978-3-319-54190-7_30 . Please refer to any applicable terms of use of the publisher.
Accepted author manuscript, 3.38 MB

Handle.net

Persistent link

Cite this

Panphattarasap, P., & Calway, A. (2017). Visual place recognition using landmark distribution descriptors. In Computer Vision - ACCV 2016: 13th Asian Conference on Computer Vision, ACCV 2016, Revised Selected Papers (Vol. 10114 LNCS, pp. 487-502). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10114 LNCS). Springer-Verlag Berlin. https://doi.org/10.1007/978-3-319-54190-7_30

Panphattarasap, Pilailuck ; Calway, Andrew. / Visual place recognition using landmark distribution descriptors. Computer Vision - ACCV 2016: 13th Asian Conference on Computer Vision, ACCV 2016, Revised Selected Papers. Vol. 10114 LNCS Springer-Verlag Berlin, 2017. pp. 487-502 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{106438e6bd7b4bbaaa7ea205e34f9db0,

title = "Visual place recognition using landmark distribution descriptors",

abstract = "Recent work by S{\"u}nderhauf et al. [1] demonstrated improved visual place recognition using proposal regions coupled with features from convolutional neural networks (CNN) to match landmarks between views. In this work we extend the approach by introducing descriptors built from landmark features which also encode the spatial distribution of the landmarks within a view. Matching descriptors then enforces consistency of the relative positions of landmarks between views. This has a significant impact on performance. For example, in experiments on 10 image-pair datasets, each consisting of 200 urban locations with significant differences in viewing positions and conditions, we recorded average precision of around 70% (at 100% recall), compared with 58% obtained using whole image CNN features and 50% for the method in [1].",

author = "Pilailuck Panphattarasap and Andrew Calway",

year = "2017",

doi = "10.1007/978-3-319-54190-7_30",

language = "English",

isbn = "9783319541891",

volume = "10114 LNCS",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer-Verlag Berlin",

pages = "487--502",

booktitle = "Computer Vision - ACCV 2016",

note = "13th Asian Conference on Computer Vision 2016 : Workshop on Assistive Vision, ACCV 16 ; Conference date: 20-11-2016 Through 24-11-2016",

url = "http://www.accv2016.org/",

}

Panphattarasap, P & Calway, A 2017, Visual place recognition using landmark distribution descriptors. in Computer Vision - ACCV 2016: 13th Asian Conference on Computer Vision, ACCV 2016, Revised Selected Papers. vol. 10114 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10114 LNCS, Springer-Verlag Berlin, pp. 487-502, 13th Asian Conference on Computer Vision 2016, Taipei, Taiwan, 20/11/16. https://doi.org/10.1007/978-3-319-54190-7_30

Visual place recognition using landmark distribution descriptors. / Panphattarasap, Pilailuck; Calway, Andrew.
Computer Vision - ACCV 2016: 13th Asian Conference on Computer Vision, ACCV 2016, Revised Selected Papers. Vol. 10114 LNCS Springer-Verlag Berlin, 2017. p. 487-502 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10114 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - Visual place recognition using landmark distribution descriptors

AU - Panphattarasap, Pilailuck

AU - Calway, Andrew

N1 - Conference code: 13

PY - 2017

Y1 - 2017

N2 - Recent work by Sünderhauf et al. [1] demonstrated improved visual place recognition using proposal regions coupled with features from convolutional neural networks (CNN) to match landmarks between views. In this work we extend the approach by introducing descriptors built from landmark features which also encode the spatial distribution of the landmarks within a view. Matching descriptors then enforces consistency of the relative positions of landmarks between views. This has a significant impact on performance. For example, in experiments on 10 image-pair datasets, each consisting of 200 urban locations with significant differences in viewing positions and conditions, we recorded average precision of around 70% (at 100% recall), compared with 58% obtained using whole image CNN features and 50% for the method in [1].

AB - Recent work by Sünderhauf et al. [1] demonstrated improved visual place recognition using proposal regions coupled with features from convolutional neural networks (CNN) to match landmarks between views. In this work we extend the approach by introducing descriptors built from landmark features which also encode the spatial distribution of the landmarks within a view. Matching descriptors then enforces consistency of the relative positions of landmarks between views. This has a significant impact on performance. For example, in experiments on 10 image-pair datasets, each consisting of 200 urban locations with significant differences in viewing positions and conditions, we recorded average precision of around 70% (at 100% recall), compared with 58% obtained using whole image CNN features and 50% for the method in [1].

UR - http://www.scopus.com/inward/record.url?scp=85016002487&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-54190-7_30

DO - 10.1007/978-3-319-54190-7_30

M3 - Conference Contribution (Conference Proceeding)

AN - SCOPUS:85016002487

SN - 9783319541891

VL - 10114 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 487

EP - 502

BT - Computer Vision - ACCV 2016

PB - Springer-Verlag Berlin

T2 - 13th Asian Conference on Computer Vision 2016

Y2 - 20 November 2016 through 24 November 2016

ER -

Panphattarasap P, Calway A. Visual place recognition using landmark distribution descriptors. In Computer Vision - ACCV 2016: 13th Asian Conference on Computer Vision, ACCV 2016, Revised Selected Papers. Vol. 10114 LNCS. Springer-Verlag Berlin. 2017. p. 487-502. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Epub 2017 Mar 12. doi: 10.1007/978-3-319-54190-7_30

Visual place recognition using landmark distribution descriptors

Abstract

Publication series

Conference

Access to Document

Handle.net

Other files and links

Fingerprint