Relative Camera Pose Estimation Using Convolutional Neural Networks

Iaroslav Melekhov; Juha Ylioinas; Juho Kannala; Esa Rahtu

doi:10.1007/978-3-319-70353-4_57

Relative Camera Pose Estimation Using Convolutional Neural Networks

Iaroslav Melekhov, Juha Ylioinas, Juho Kannala, Esa Rahtu

Tampere University of Technology

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

125 Sitaatiot (Scopus)

Abstrakti

This paper presents a convolutional neural network based approach for estimating the relative pose between two cameras. The proposed network takes RGB images from both cameras as input and directly produces the relative rotation and translation as output. The system is trained in an end-to-end manner utilising transfer learning from a large scale classification dataset. The introduced approach is compared with widely used local feature based methods (SURF, ORB) and the results indicate a clear improvement over the baseline. In addition, a variant of the proposed architecture containing a spatial pyramid pooling (SPP) layer is evaluated and shown to further improve the performance.

Alkuperäiskieli	Englanti
Otsikko	Advanced Concepts for Intelligent Vision Systems
Alaotsikko	18th International Conference, ACIVS 2017, Antwerp, Belgium, September 18-21, 2017, Proceedings
Toimittajat	Jacques Blanc-Talon, Rudi Penne, Wilfried Philips, Dan Popescu, Paul Scheunders
Julkaisupaikka	Cham
Kustantaja	Springer
Sivut	675-687
Sivumäärä	13
ISBN (elektroninen)	978-3-319-70353-4
ISBN (painettu)	978-3-319-70352-7
DOI - pysyväislinkit	https://doi.org/10.1007/978-3-319-70353-4_57
Tila	Julkaistu - 2017
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	International Conference on Advanced Concepts for Intelligent Vision Systems - Antwerp, Belgia Kesto: 18 syysk. 2017 → 21 syysk. 2017 Konferenssinumero: 18

Julkaisusarja

Nimi	Lecture Notes in Computer Science
Kustantaja	Springer
Vuosikerta	10617
ISSN (painettu)	0302-9743
ISSN (elektroninen)	1611-3349

Conference

Conference	International Conference on Advanced Concepts for Intelligent Vision Systems
Lyhennettä	ACIVS
Maa/Alue	Belgia
Kaupunki	Antwerp
Ajanjakso	18/09/2017 → 21/09/2017

Pääsy asiakirjaan

10.1007/978-3-319-70353-4_57

https://arxiv.org/abs/1702.01381

Muut tiedostot ja linkit

Link to publication in Scopus

Siteeraa tätä

Melekhov, I., Ylioinas, J., Kannala, J., & Rahtu, E. (2017). Relative Camera Pose Estimation Using Convolutional Neural Networks. teoksessa J. Blanc-Talon, R. Penne, W. Philips, D. Popescu, & P. Scheunders (Toimittajat), Advanced Concepts for Intelligent Vision Systems: 18th International Conference, ACIVS 2017, Antwerp, Belgium, September 18-21, 2017, Proceedings (Sivut 675-687). (Lecture Notes in Computer Science; Vuosikerta 10617). Springer. https://doi.org/10.1007/978-3-319-70353-4_57

Melekhov, Iaroslav ; Ylioinas, Juha ; Kannala, Juho et al. / Relative Camera Pose Estimation Using Convolutional Neural Networks. Advanced Concepts for Intelligent Vision Systems: 18th International Conference, ACIVS 2017, Antwerp, Belgium, September 18-21, 2017, Proceedings. Toimittaja / Jacques Blanc-Talon ; Rudi Penne ; Wilfried Philips ; Dan Popescu ; Paul Scheunders. Cham : Springer, 2017. Sivut 675-687 (Lecture Notes in Computer Science).

@inproceedings{4c9a92bfe7c74c16b2e355c4d130dfb6,

title = "Relative Camera Pose Estimation Using Convolutional Neural Networks",

abstract = "This paper presents a convolutional neural network based approach for estimating the relative pose between two cameras. The proposed network takes RGB images from both cameras as input and directly produces the relative rotation and translation as output. The system is trained in an end-to-end manner utilising transfer learning from a large scale classification dataset. The introduced approach is compared with widely used local feature based methods (SURF, ORB) and the results indicate a clear improvement over the baseline. In addition, a variant of the proposed architecture containing a spatial pyramid pooling (SPP) layer is evaluated and shown to further improve the performance.",

author = "Iaroslav Melekhov and Juha Ylioinas and Juho Kannala and Esa Rahtu",

year = "2017",

doi = "10.1007/978-3-319-70353-4\_57",

language = "English",

isbn = "978-3-319-70352-7",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "675--687",

editor = "Jacques Blanc-Talon and Rudi Penne and Wilfried Philips and Dan Popescu and Paul Scheunders",

booktitle = "Advanced Concepts for Intelligent Vision Systems",

address = "Germany",

note = "International Conference on Advanced Concepts for Intelligent Vision Systems, ACIVS ; Conference date: 18-09-2017 Through 21-09-2017",

}

Melekhov, I, Ylioinas, J, Kannala, J & Rahtu, E 2017, Relative Camera Pose Estimation Using Convolutional Neural Networks. julkaisussa J Blanc-Talon, R Penne, W Philips, D Popescu & P Scheunders (toim), Advanced Concepts for Intelligent Vision Systems: 18th International Conference, ACIVS 2017, Antwerp, Belgium, September 18-21, 2017, Proceedings. Lecture Notes in Computer Science, Vuosikerta 10617, Springer, Cham, Sivut 675-687, International Conference on Advanced Concepts for Intelligent Vision Systems, Antwerp, Belgia, 18/09/2017. https://doi.org/10.1007/978-3-319-70353-4_57

Relative Camera Pose Estimation Using Convolutional Neural Networks. / Melekhov, Iaroslav; Ylioinas, Juha; Kannala, Juho et al.
Advanced Concepts for Intelligent Vision Systems: 18th International Conference, ACIVS 2017, Antwerp, Belgium, September 18-21, 2017, Proceedings. toim. / Jacques Blanc-Talon; Rudi Penne; Wilfried Philips; Dan Popescu; Paul Scheunders. Cham: Springer, 2017. s. 675-687 (Lecture Notes in Computer Science; Vuosikerta 10617).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

TY - GEN

T1 - Relative Camera Pose Estimation Using Convolutional Neural Networks

AU - Melekhov, Iaroslav

AU - Ylioinas, Juha

AU - Kannala, Juho

AU - Rahtu, Esa

N1 - Conference code: 18

PY - 2017

Y1 - 2017

N2 - This paper presents a convolutional neural network based approach for estimating the relative pose between two cameras. The proposed network takes RGB images from both cameras as input and directly produces the relative rotation and translation as output. The system is trained in an end-to-end manner utilising transfer learning from a large scale classification dataset. The introduced approach is compared with widely used local feature based methods (SURF, ORB) and the results indicate a clear improvement over the baseline. In addition, a variant of the proposed architecture containing a spatial pyramid pooling (SPP) layer is evaluated and shown to further improve the performance.

AB - This paper presents a convolutional neural network based approach for estimating the relative pose between two cameras. The proposed network takes RGB images from both cameras as input and directly produces the relative rotation and translation as output. The system is trained in an end-to-end manner utilising transfer learning from a large scale classification dataset. The introduced approach is compared with widely used local feature based methods (SURF, ORB) and the results indicate a clear improvement over the baseline. In addition, a variant of the proposed architecture containing a spatial pyramid pooling (SPP) layer is evaluated and shown to further improve the performance.

UR - https://www.scopus.com/pages/publications/85036668143

U2 - 10.1007/978-3-319-70353-4_57

DO - 10.1007/978-3-319-70353-4_57

M3 - Conference article in proceedings

SN - 978-3-319-70352-7

T3 - Lecture Notes in Computer Science

SP - 675

EP - 687

BT - Advanced Concepts for Intelligent Vision Systems

A2 - Blanc-Talon, Jacques

A2 - Penne, Rudi

A2 - Philips, Wilfried

A2 - Popescu, Dan

A2 - Scheunders, Paul

PB - Springer

CY - Cham

T2 - International Conference on Advanced Concepts for Intelligent Vision Systems

Y2 - 18 September 2017 through 21 September 2017

ER -

Melekhov I, Ylioinas J, Kannala J, Rahtu E. Relative Camera Pose Estimation Using Convolutional Neural Networks. julkaisussa Blanc-Talon J, Penne R, Philips W, Popescu D, Scheunders P, toimittajat, Advanced Concepts for Intelligent Vision Systems: 18th International Conference, ACIVS 2017, Antwerp, Belgium, September 18-21, 2017, Proceedings. Cham: Springer. 2017. s. 675-687. (Lecture Notes in Computer Science). doi: 10.1007/978-3-319-70353-4_57

Relative Camera Pose Estimation Using Convolutional Neural Networks

Abstrakti

Julkaisusarja

Conference

Pääsy asiakirjaan

Muut tiedostot ja linkit

Sormenjälki

Siteeraa tätä