Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection

Florian Kluger; Hanno Ackermann; Michael Ying Yang; Bodo Rosenhahn

doi:10.1007/978-3-319-66709-6_2

Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection

Florian Kluger
, Hanno Ackermann
, Michael Ying Yang
, Bodo Rosenhahn

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

31 Citations (Scopus)

41 Downloads (Pure)

Abstract

We present a novel approach for vanishing point detection from uncalibrated monocular images. In contrast to state-of-the-art, we make no a priori assumptions about the observed scene. Our method is based on a convolutional neural network (CNN) which does not use natural images, but a Gaussian sphere representation arising from an inverse gnomonic projection of lines detected in an image. This allows us to rely on synthetic data for training, eliminating the need for labelled images. Our method achieves competitive performance on three horizon estimation benchmark datasets. We further highlight some additional use cases for which our vanishing point detection algorithm can be used.

Original language	English
Title of host publication	Pattern recognition : Proceedings of the 39th German Conference, GCPR 2017, Basel, Switzerland, 12-15 September, 2017
Publisher	Springer
Pages	17-28
DOIs	https://doi.org/10.1007/978-3-319-66709-6_2
Publication status	Published - 15 Aug 2017
Event	39th German Conference on Pattern Recognition, GCPR 2017 - Basel, Switzerland Duration: 12 Sept 2017 → 15 Sept 2017 Conference number: 39

Publication series

Name	Lecture Notes in Computer Science
Volume	10496

Conference

Conference	39th German Conference on Pattern Recognition, GCPR 2017
Abbreviated title	GCPR 2017
Country/Territory	Switzerland
City	Basel
Period	12/09/17 → 15/09/17

Keywords

2025 OA procedure

Access to Document

10.1007/978-3-319-66709-6_2

978-3-319-66709-6_2Final published version, 1.75 MBLicence: Taverne

Cite this

Kluger, F., Ackermann, H., Yang, M. Y., & Rosenhahn, B. (2017). Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection. In Pattern recognition : Proceedings of the 39th German Conference, GCPR 2017, Basel, Switzerland, 12-15 September, 2017 (pp. 17-28). Article Chapter 2 (Lecture Notes in Computer Science; Vol. 10496). Springer. https://doi.org/10.1007/978-3-319-66709-6_2

@inbook{f27566f81a9d4806a48fcd93df66717a,

title = "Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection",

abstract = "We present a novel approach for vanishing point detection from uncalibrated monocular images. In contrast to state-of-the-art, we make no a priori assumptions about the observed scene. Our method is based on a convolutional neural network (CNN) which does not use natural images, but a Gaussian sphere representation arising from an inverse gnomonic projection of lines detected in an image. This allows us to rely on synthetic data for training, eliminating the need for labelled images. Our method achieves competitive performance on three horizon estimation benchmark datasets. We further highlight some additional use cases for which our vanishing point detection algorithm can be used.",

keywords = "2025 OA procedure",

author = "Florian Kluger and Hanno Ackermann and Yang, \{Michael Ying\} and Bodo Rosenhahn",

year = "2017",

month = aug,

day = "15",

doi = "10.1007/978-3-319-66709-6\_2",

language = "English",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "17--28",

booktitle = "Pattern recognition : Proceedings of the 39th German Conference, GCPR 2017, Basel, Switzerland, 12-15 September, 2017",

address = "Germany",

note = "39th German Conference on Pattern Recognition, GCPR 2017, GCPR 2017 ; Conference date: 12-09-2017 Through 15-09-2017",

}

Kluger, F, Ackermann, H, Yang, MY & Rosenhahn, B 2017, Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection. in Pattern recognition : Proceedings of the 39th German Conference, GCPR 2017, Basel, Switzerland, 12-15 September, 2017., Chapter 2, Lecture Notes in Computer Science, vol. 10496, Springer, pp. 17-28, 39th German Conference on Pattern Recognition, GCPR 2017, Basel, Switzerland, 12/09/17. https://doi.org/10.1007/978-3-319-66709-6_2

Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection. / Kluger, Florian; Ackermann, Hanno; Yang, Michael Ying et al.
Pattern recognition : Proceedings of the 39th German Conference, GCPR 2017, Basel, Switzerland, 12-15 September, 2017. Springer, 2017. p. 17-28 Chapter 2 (Lecture Notes in Computer Science; Vol. 10496).

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

TY - CHAP

T1 - Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection

AU - Kluger, Florian

AU - Ackermann, Hanno

AU - Yang, Michael Ying

AU - Rosenhahn, Bodo

N1 - Conference code: 39

PY - 2017/8/15

Y1 - 2017/8/15

N2 - We present a novel approach for vanishing point detection from uncalibrated monocular images. In contrast to state-of-the-art, we make no a priori assumptions about the observed scene. Our method is based on a convolutional neural network (CNN) which does not use natural images, but a Gaussian sphere representation arising from an inverse gnomonic projection of lines detected in an image. This allows us to rely on synthetic data for training, eliminating the need for labelled images. Our method achieves competitive performance on three horizon estimation benchmark datasets. We further highlight some additional use cases for which our vanishing point detection algorithm can be used.

AB - We present a novel approach for vanishing point detection from uncalibrated monocular images. In contrast to state-of-the-art, we make no a priori assumptions about the observed scene. Our method is based on a convolutional neural network (CNN) which does not use natural images, but a Gaussian sphere representation arising from an inverse gnomonic projection of lines detected in an image. This allows us to rely on synthetic data for training, eliminating the need for labelled images. Our method achieves competitive performance on three horizon estimation benchmark datasets. We further highlight some additional use cases for which our vanishing point detection algorithm can be used.

KW - 2025 OA procedure

U2 - 10.1007/978-3-319-66709-6_2

DO - 10.1007/978-3-319-66709-6_2

M3 - Chapter

T3 - Lecture Notes in Computer Science

SP - 17

EP - 28

BT - Pattern recognition : Proceedings of the 39th German Conference, GCPR 2017, Basel, Switzerland, 12-15 September, 2017

PB - Springer

T2 - 39th German Conference on Pattern Recognition, GCPR 2017

Y2 - 12 September 2017 through 15 September 2017

ER -

Deep Learning for Vanishing Point Detection Using an Inverse Gnomonic Projection

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this