Fast 3D scene alignment with stereo images using a Stixel-based 3D model

Dennis W.J.M. Van De Wouw; Willem P. Sanberg; Gijs Dubbelman; Peter H.N. De With

doi:10.5220/0006535302500259

Fast 3D scene alignment with stereo images using a Stixel-based 3D model

Dennis W.J.M. Van De Wouw, Willem P. Sanberg, Gijs Dubbelman, Peter H.N. De With

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

3 Citations (Scopus)

4 Downloads (Pure)

Abstract

Scene alignment for images recorded from different viewpoints is a challenging task, especially considering strong parallax effects. This work proposes a diorama-box model for a 2.5D hierarchical alignment approach, which is specifically designed for image registration from a moving vehicle using a stereo camera. For this purpose, the Stixel World algorithm is used to partition the scene into super-pixels, which are transformed to 3D. This model is further refined by assigning a slanting orientation to each stixel and by interpolating between stixels, to prevent gaps in the 3D model. The resulting alignment shows promising results, where under normal viewing conditions, more than 96% of all annotated points are registered with an alignment error up to 5 pixels at a resolution of 1920 × 1440 pixels, executing at near-real time performance (4 fps) for the intended application.

Original language	English
Title of host publication	Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
Editors	Alain Tremeau, Francisco Imai, Jose Braz
Place of Publication	Sétubal
Publisher	SciTePress Digital Library
Pages	250-259
Number of pages	10
Volume	4
ISBN (Print)	978-989-758-290-5
DOIs	https://doi.org/10.5220/0006535302500259
Publication status	Published - 27 Jan 2018
Event	13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2018) - Funchal, Madeira, Portugal Duration: 27 Jan 2018 → 29 Jan 2018 Conference number: 13

Conference

Conference	13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2018)
Abbreviated title	VISIGRAPP 2018
Country/Territory	Portugal
City	Funchal, Madeira
Period	27/01/18 → 29/01/18

Keywords

Image registration
Real-time
Stixel-world
View synthesis

Access to Document

10.5220/0006535302500259

Cite this

Van De Wouw, D. W. J. M., Sanberg, W. P., Dubbelman, G., & De With, P. H. N. (2018). Fast 3D scene alignment with stereo images using a Stixel-based 3D model. In A. Tremeau, F. Imai, & J. Braz (Eds.), Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (Vol. 4, pp. 250-259). SciTePress Digital Library. https://doi.org/10.5220/0006535302500259

Van De Wouw, Dennis W.J.M. ; Sanberg, Willem P. ; Dubbelman, Gijs et al. / Fast 3D scene alignment with stereo images using a Stixel-based 3D model. Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. editor / Alain Tremeau ; Francisco Imai ; Jose Braz. Vol. 4 Sétubal : SciTePress Digital Library, 2018. pp. 250-259

@inproceedings{1fa05273624a400e9465f83400cb24b3,

title = "Fast 3D scene alignment with stereo images using a Stixel-based 3D model",

abstract = "Scene alignment for images recorded from different viewpoints is a challenging task, especially considering strong parallax effects. This work proposes a diorama-box model for a 2.5D hierarchical alignment approach, which is specifically designed for image registration from a moving vehicle using a stereo camera. For this purpose, the Stixel World algorithm is used to partition the scene into super-pixels, which are transformed to 3D. This model is further refined by assigning a slanting orientation to each stixel and by interpolating between stixels, to prevent gaps in the 3D model. The resulting alignment shows promising results, where under normal viewing conditions, more than 96% of all annotated points are registered with an alignment error up to 5 pixels at a resolution of 1920 × 1440 pixels, executing at near-real time performance (4 fps) for the intended application.",

keywords = "Image registration, Real-time, Stixel-world, View synthesis",

author = "{Van De Wouw}, {Dennis W.J.M.} and Sanberg, {Willem P.} and Gijs Dubbelman and {De With}, {Peter H.N.}",

year = "2018",

month = jan,

day = "27",

doi = "10.5220/0006535302500259",

language = "English",

isbn = "978-989-758-290-5 ",

volume = "4",

pages = "250--259",

editor = "Alain Tremeau and Francisco Imai and Jose Braz",

booktitle = "Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications",

publisher = "SciTePress Digital Library",

note = "13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2018), VISIGRAPP 2018 ; Conference date: 27-01-2018 Through 29-01-2018",

}

Van De Wouw, DWJM, Sanberg, WP, Dubbelman, G & De With, PHN 2018, Fast 3D scene alignment with stereo images using a Stixel-based 3D model. in A Tremeau, F Imai & J Braz (eds), Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. vol. 4, SciTePress Digital Library, Sétubal, pp. 250-259, 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2018), Funchal, Madeira, Portugal, 27/01/18. https://doi.org/10.5220/0006535302500259

Fast 3D scene alignment with stereo images using a Stixel-based 3D model. / Van De Wouw, Dennis W.J.M.; Sanberg, Willem P.; Dubbelman, Gijs et al.
Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. ed. / Alain Tremeau; Francisco Imai; Jose Braz. Vol. 4 Sétubal: SciTePress Digital Library, 2018. p. 250-259.

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

TY - GEN

T1 - Fast 3D scene alignment with stereo images using a Stixel-based 3D model

AU - Van De Wouw, Dennis W.J.M.

AU - Sanberg, Willem P.

AU - Dubbelman, Gijs

AU - De With, Peter H.N.

N1 - Conference code: 13

PY - 2018/1/27

Y1 - 2018/1/27

N2 - Scene alignment for images recorded from different viewpoints is a challenging task, especially considering strong parallax effects. This work proposes a diorama-box model for a 2.5D hierarchical alignment approach, which is specifically designed for image registration from a moving vehicle using a stereo camera. For this purpose, the Stixel World algorithm is used to partition the scene into super-pixels, which are transformed to 3D. This model is further refined by assigning a slanting orientation to each stixel and by interpolating between stixels, to prevent gaps in the 3D model. The resulting alignment shows promising results, where under normal viewing conditions, more than 96% of all annotated points are registered with an alignment error up to 5 pixels at a resolution of 1920 × 1440 pixels, executing at near-real time performance (4 fps) for the intended application.

AB - Scene alignment for images recorded from different viewpoints is a challenging task, especially considering strong parallax effects. This work proposes a diorama-box model for a 2.5D hierarchical alignment approach, which is specifically designed for image registration from a moving vehicle using a stereo camera. For this purpose, the Stixel World algorithm is used to partition the scene into super-pixels, which are transformed to 3D. This model is further refined by assigning a slanting orientation to each stixel and by interpolating between stixels, to prevent gaps in the 3D model. The resulting alignment shows promising results, where under normal viewing conditions, more than 96% of all annotated points are registered with an alignment error up to 5 pixels at a resolution of 1920 × 1440 pixels, executing at near-real time performance (4 fps) for the intended application.

KW - Image registration

KW - Real-time

KW - Stixel-world

KW - View synthesis

UR - http://www.scopus.com/inward/record.url?scp=85047850171&partnerID=8YFLogxK

U2 - 10.5220/0006535302500259

DO - 10.5220/0006535302500259

M3 - Conference contribution

AN - SCOPUS:85047850171

SN - 978-989-758-290-5

VL - 4

SP - 250

EP - 259

BT - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

A2 - Tremeau, Alain

A2 - Imai, Francisco

A2 - Braz, Jose

PB - SciTePress Digital Library

CY - Sétubal

T2 - 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2018)

Y2 - 27 January 2018 through 29 January 2018

ER -

Van De Wouw DWJM, Sanberg WP, Dubbelman G , De With PHN. Fast 3D scene alignment with stereo images using a Stixel-based 3D model. In Tremeau A, Imai F, Braz J, editors, Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Vol. 4. Sétubal: SciTePress Digital Library. 2018. p. 250-259 doi: 10.5220/0006535302500259

Fast 3D scene alignment with stereo images using a Stixel-based 3D model

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this