Geometric attention for prediction of differential properties in 3d point clouds

Albert Matveev; Alexey Artemov; Denis Zorin; Evgeny Burnaev

doi:10.1007/978-3-030-58309-5_9

Geometric attention for prediction of differential properties in 3d point clouds

Albert Matveev, Alexey Artemov, Denis Zorin, Evgeny Burnaev

Computer Science

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

Abstract

Estimation of differential geometric quantities in discrete 3D data representations is one of the crucial steps in the geometry processing pipeline. Specifically, estimating normals and sharp feature lines from raw point clouds helps improve meshing quality and allows us to use more precise surface reconstruction techniques. When designing a learnable approach to such problems, the main difficulty is selecting neighborhoods in a point cloud and incorporating geometric relations between the points. In this study, we present a geometric attention mechanism that can provide such properties in a learnable fashion. We establish the usefulness of the proposed technique with several experiments on the prediction of normal vectors and the extraction of feature lines.

Original language	English (US)
Title of host publication	Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings
Editors	Frank-Peter Schilling, Thilo Stadelmann
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	113-124
Number of pages	12
ISBN (Print)	9783030583088
DOIs	https://doi.org/10.1007/978-3-030-58309-5_9
State	Published - 2020
Event	9th IAPR TC3 International Workshop on Artificial Neural Networks in Pattern Recognition, ANNPR 2020 - Winterthur, Switzerland Duration: Sep 2 2020 → Sep 4 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12294 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	9th IAPR TC3 International Workshop on Artificial Neural Networks in Pattern Recognition, ANNPR 2020
Country/Territory	Switzerland
City	Winterthur
Period	9/2/20 → 9/4/20

Keywords

3D computer vision
3D point clouds
Attention

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-030-58309-5_9

Cite this

Matveev, A., Artemov, A., Zorin, D., & Burnaev, E. (2020). Geometric attention for prediction of differential properties in 3d point clouds. In F-P. Schilling, & T. Stadelmann (Eds.), Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings (pp. 113-124). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12294 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-58309-5_9

Geometric attention for prediction of differential properties in 3d point clouds. / Matveev, Albert; Artemov, Alexey; Zorin, Denis et al.

Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings. ed. / Frank-Peter Schilling; Thilo Stadelmann. Springer Science and Business Media Deutschland GmbH, 2020. p. 113-124 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12294 LNAI).

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

Matveev, A, Artemov, A, Zorin, D & Burnaev, E 2020, Geometric attention for prediction of differential properties in 3d point clouds. in F-P Schilling & T Stadelmann (eds), Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12294 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 113-124, 9th IAPR TC3 International Workshop on Artificial Neural Networks in Pattern Recognition, ANNPR 2020, Winterthur, Switzerland, 9/2/20. https://doi.org/10.1007/978-3-030-58309-5_9

Matveev A, Artemov A, Zorin D, Burnaev E. Geometric attention for prediction of differential properties in 3d point clouds. In Schilling F-P, Stadelmann T, editors, Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 113-124. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-58309-5_9

Matveev, Albert ; Artemov, Alexey ; Zorin, Denis et al. / Geometric attention for prediction of differential properties in 3d point clouds. Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings. editor / Frank-Peter Schilling ; Thilo Stadelmann. Springer Science and Business Media Deutschland GmbH, 2020. pp. 113-124 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{5a873f0da91549f984376850dc4e0236,

title = "Geometric attention for prediction of differential properties in 3d point clouds",

abstract = "Estimation of differential geometric quantities in discrete 3D data representations is one of the crucial steps in the geometry processing pipeline. Specifically, estimating normals and sharp feature lines from raw point clouds helps improve meshing quality and allows us to use more precise surface reconstruction techniques. When designing a learnable approach to such problems, the main difficulty is selecting neighborhoods in a point cloud and incorporating geometric relations between the points. In this study, we present a geometric attention mechanism that can provide such properties in a learnable fashion. We establish the usefulness of the proposed technique with several experiments on the prediction of normal vectors and the extraction of feature lines.",

keywords = "3D computer vision, 3D point clouds, Attention",

author = "Albert Matveev and Alexey Artemov and Denis Zorin and Evgeny Burnaev",

note = "Funding Information: Acknowledgments. The reported study was funded by RFBR, project number 19-31-90144. The Authors acknowledge the usage of the Skoltech CDISE HPC cluster Zhores for obtaining the results presented in this paper. Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2020.; 9th IAPR TC3 International Workshop on Artificial Neural Networks in Pattern Recognition, ANNPR 2020 ; Conference date: 02-09-2020 Through 04-09-2020",

year = "2020",

doi = "10.1007/978-3-030-58309-5_9",

language = "English (US)",

isbn = "9783030583088",

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 = "113--124",

editor = "Frank-Peter Schilling and Thilo Stadelmann",

booktitle = "Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings",

address = "Germany",

}

TY - GEN

T1 - Geometric attention for prediction of differential properties in 3d point clouds

AU - Matveev, Albert

AU - Artemov, Alexey

AU - Zorin, Denis

AU - Burnaev, Evgeny

N1 - Funding Information: Acknowledgments. The reported study was funded by RFBR, project number 19-31-90144. The Authors acknowledge the usage of the Skoltech CDISE HPC cluster Zhores for obtaining the results presented in this paper. Publisher Copyright: © Springer Nature Switzerland AG 2020.

PY - 2020

Y1 - 2020

N2 - Estimation of differential geometric quantities in discrete 3D data representations is one of the crucial steps in the geometry processing pipeline. Specifically, estimating normals and sharp feature lines from raw point clouds helps improve meshing quality and allows us to use more precise surface reconstruction techniques. When designing a learnable approach to such problems, the main difficulty is selecting neighborhoods in a point cloud and incorporating geometric relations between the points. In this study, we present a geometric attention mechanism that can provide such properties in a learnable fashion. We establish the usefulness of the proposed technique with several experiments on the prediction of normal vectors and the extraction of feature lines.

AB - Estimation of differential geometric quantities in discrete 3D data representations is one of the crucial steps in the geometry processing pipeline. Specifically, estimating normals and sharp feature lines from raw point clouds helps improve meshing quality and allows us to use more precise surface reconstruction techniques. When designing a learnable approach to such problems, the main difficulty is selecting neighborhoods in a point cloud and incorporating geometric relations between the points. In this study, we present a geometric attention mechanism that can provide such properties in a learnable fashion. We establish the usefulness of the proposed technique with several experiments on the prediction of normal vectors and the extraction of feature lines.

KW - 3D computer vision

KW - 3D point clouds

KW - Attention

UR - http://www.scopus.com/inward/record.url?scp=85091281933&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85091281933&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-58309-5_9

DO - 10.1007/978-3-030-58309-5_9

M3 - Conference contribution

AN - SCOPUS:85091281933

SN - 9783030583088

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

SP - 113

EP - 124

BT - Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Proceedings

A2 - Schilling, Frank-Peter

A2 - Stadelmann, Thilo

PB - Springer Science and Business Media Deutschland GmbH

T2 - 9th IAPR TC3 International Workshop on Artificial Neural Networks in Pattern Recognition, ANNPR 2020

Y2 - 2 September 2020 through 4 September 2020

ER -

Geometric attention for prediction of differential properties in 3d point clouds

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this