Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application

Xuchong Qiu; Yang Xiao; Chaohui Wang; Renaud Marlet

doi:10.1007/978-3-030-58548-8_40

Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application

Xuchong Qiu
, Yang Xiao
, Chaohui Wang
, Renaud Marlet

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

Abstract

We formalize concepts around geometric occlusion in 2D images (i.e., ignoring semantics), and propose a novel unified formulation of both occlusion boundaries and occlusion orientations via a pixel-pair occlusion relation. The former provides a way to generate large-scale accurate occlusion datasets while, based on the latter, we propose a novel method for task-independent pixel-level occlusion relationship estimation from single images. Experiments on a variety of datasets demonstrate that our method outperforms existing ones on this task. To further illustrate the value of our formulation, we also propose a new depth map refinement method that consistently improve the performance of state-of-the-art monocular depth estimation methods.

Original language	English
Title of host publication	Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings
Editors	Andrea Vedaldi, Horst Bischof, Thomas Brox, Jan-Michael Frahm
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	690-708
Number of pages	19
ISBN (Print)	9783030585471
DOIs	https://doi.org/10.1007/978-3-030-58548-8_40
Publication status	Published - 1 Jan 2020
Externally published	Yes
Event	16th European Conference on Computer Vision, ECCV 2020 - Glasgow, United Kingdom Duration: 23 Aug 2020 → 28 Aug 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12349 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	16th European Conference on Computer Vision, ECCV 2020
Country/Territory	United Kingdom
City	Glasgow
Period	23/08/20 → 28/08/20

Keywords

Depth refinement
Occlusion boundary
Occlusion relation

Access to Document

10.1007/978-3-030-58548-8_40

Cite this

Qiu, X., Xiao, Y., Wang, C., & Marlet, R. (2020). Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application. In A. Vedaldi, H. Bischof, T. Brox, & J.-M. Frahm (Eds.), Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings (pp. 690-708). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12349 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-58548-8_40

Qiu, Xuchong ; Xiao, Yang ; Wang, Chaohui et al. / Pixel-Pair Occlusion Relationship Map (P2ORM) : Formulation, Inference and Application. Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings. editor / Andrea Vedaldi ; Horst Bischof ; Thomas Brox ; Jan-Michael Frahm. Springer Science and Business Media Deutschland GmbH, 2020. pp. 690-708 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{aa0a7aff19444b55aa082eb9909265df,

title = "Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application",

abstract = "We formalize concepts around geometric occlusion in 2D images (i.e., ignoring semantics), and propose a novel unified formulation of both occlusion boundaries and occlusion orientations via a pixel-pair occlusion relation. The former provides a way to generate large-scale accurate occlusion datasets while, based on the latter, we propose a novel method for task-independent pixel-level occlusion relationship estimation from single images. Experiments on a variety of datasets demonstrate that our method outperforms existing ones on this task. To further illustrate the value of our formulation, we also propose a new depth map refinement method that consistently improve the performance of state-of-the-art monocular depth estimation methods.",

keywords = "Depth refinement, Occlusion boundary, Occlusion relation",

author = "Xuchong Qiu and Yang Xiao and Chaohui Wang and Renaud Marlet",

note = "Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.; 16th European Conference on Computer Vision, ECCV 2020 ; Conference date: 23-08-2020 Through 28-08-2020",

year = "2020",

month = jan,

day = "1",

doi = "10.1007/978-3-030-58548-8\_40",

language = "English",

isbn = "9783030585471",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "690--708",

editor = "Andrea Vedaldi and Horst Bischof and Thomas Brox and Jan-Michael Frahm",

booktitle = "Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings",

}

Qiu, X, Xiao, Y, Wang, C & Marlet, R 2020, Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application. in A Vedaldi, H Bischof, T Brox & J-M Frahm (eds), Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12349 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 690-708, 16th European Conference on Computer Vision, ECCV 2020, Glasgow, United Kingdom, 23/08/20. https://doi.org/10.1007/978-3-030-58548-8_40

Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application. / Qiu, Xuchong; Xiao, Yang; Wang, Chaohui et al.
Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings. ed. / Andrea Vedaldi; Horst Bischof; Thomas Brox; Jan-Michael Frahm. Springer Science and Business Media Deutschland GmbH, 2020. p. 690-708 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12349 LNCS).

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

TY - GEN

T1 - Pixel-Pair Occlusion Relationship Map (P2ORM)

T2 - 16th European Conference on Computer Vision, ECCV 2020

AU - Qiu, Xuchong

AU - Xiao, Yang

AU - Wang, Chaohui

AU - Marlet, Renaud

PY - 2020/1/1

Y1 - 2020/1/1

N2 - We formalize concepts around geometric occlusion in 2D images (i.e., ignoring semantics), and propose a novel unified formulation of both occlusion boundaries and occlusion orientations via a pixel-pair occlusion relation. The former provides a way to generate large-scale accurate occlusion datasets while, based on the latter, we propose a novel method for task-independent pixel-level occlusion relationship estimation from single images. Experiments on a variety of datasets demonstrate that our method outperforms existing ones on this task. To further illustrate the value of our formulation, we also propose a new depth map refinement method that consistently improve the performance of state-of-the-art monocular depth estimation methods.

AB - We formalize concepts around geometric occlusion in 2D images (i.e., ignoring semantics), and propose a novel unified formulation of both occlusion boundaries and occlusion orientations via a pixel-pair occlusion relation. The former provides a way to generate large-scale accurate occlusion datasets while, based on the latter, we propose a novel method for task-independent pixel-level occlusion relationship estimation from single images. Experiments on a variety of datasets demonstrate that our method outperforms existing ones on this task. To further illustrate the value of our formulation, we also propose a new depth map refinement method that consistently improve the performance of state-of-the-art monocular depth estimation methods.

KW - Depth refinement

KW - Occlusion boundary

KW - Occlusion relation

UR - https://www.scopus.com/pages/publications/85097416120

U2 - 10.1007/978-3-030-58548-8_40

DO - 10.1007/978-3-030-58548-8_40

M3 - Conference contribution

AN - SCOPUS:85097416120

SN - 9783030585471

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

SP - 690

EP - 708

BT - Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings

A2 - Vedaldi, Andrea

A2 - Bischof, Horst

A2 - Brox, Thomas

A2 - Frahm, Jan-Michael

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 23 August 2020 through 28 August 2020

ER -

Qiu X, Xiao Y, Wang C, Marlet R. Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application. In Vedaldi A, Bischof H, Brox T, Frahm JM, editors, Computer Vision – ECCV 2020 - 16th European Conference, 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 690-708. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-58548-8_40

Pixel-Pair Occlusion Relationship Map (P2ORM): Formulation, Inference and Application

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this