Point Pair Feature Matching:   Evaluating Methods to Detect Simple Shapes

Markus Ziegler; Martin Rudorfer; Xaver Kroischke; Sebastian Krone; Jörg Krüger

doi:10.1007/978-3-030-34995-0_40

Point Pair Feature Matching: Evaluating Methods to Detect Simple Shapes

Markus Ziegler^*, Martin Rudorfer, Xaver Kroischke, Sebastian Krone, Jörg Krüger

^*Corresponding author for this work

Research output: Chapter in Book/Published conference output › Conference publication

Abstract

A recent benchmark for 3D object detection and 6D pose estimation from RGB-D images shows the dominance of methods based on Point Pair Feature Matching (PPFM). Since its invention in 2010 several modifications have been proposed to cope with its weaknesses, which are computational complexity, sensitivity to noise, and difficulties in the detection of geometrically simple objects with planar surfaces and rotational symmetries. In this work we focus on the latter. We present a novel approach to automatically detect rotational symmetries by matching the object model to itself. Furthermore, we adapt methods for pose verification and use more discriminative features which incorporate global information into the Point Pair Feature. We also examine the effects of other, already existing extensions by testing them on our specialized dataset for geometrically primitive objects. Results show that particularly our handling of symmetries and the augmented features are able to boost recognition rates.

Original language	English
Title of host publication	Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings
Editors	Dimitrios Tzovaras, Dimitrios Giakoumis, Markus Vincze, Antonis Argyros
Pages	445-456
Number of pages	12
DOIs	https://doi.org/10.1007/978-3-030-34995-0_40
Publication status	Published - 23 Nov 2019
Event	12th International Conference on Computer Vision Systems, ICVS 2019 - Thessaloniki, Greece Duration: 23 Sept 2019 → 25 Sept 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11754 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	12th International Conference on Computer Vision Systems, ICVS 2019
Country/Territory	Greece
City	Thessaloniki
Period	23/09/19 → 25/09/19

Bibliographical note

Publisher Copyright:
© 2019, Springer Nature Switzerland AG.

Keywords

Object detection
Point Pair Features
Pose estimation

Access to Document

10.1007/978-3-030-34995-0_40

Cite this

Ziegler, M., Rudorfer, M., Kroischke, X., Krone, S., & Krüger, J. (2019). Point Pair Feature Matching: Evaluating Methods to Detect Simple Shapes. In D. Tzovaras, D. Giakoumis, M. Vincze, & A. Argyros (Eds.), Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings (pp. 445-456). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11754 LNCS). https://doi.org/10.1007/978-3-030-34995-0_40

Ziegler, Markus ; Rudorfer, Martin ; Kroischke, Xaver et al. / Point Pair Feature Matching : Evaluating Methods to Detect Simple Shapes. Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings. editor / Dimitrios Tzovaras ; Dimitrios Giakoumis ; Markus Vincze ; Antonis Argyros. 2019. pp. 445-456 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{cd6d95c5c7f24e29b387b55e9e49d067,

title = "Point Pair Feature Matching: Evaluating Methods to Detect Simple Shapes",

abstract = "A recent benchmark for 3D object detection and 6D pose estimation from RGB-D images shows the dominance of methods based on Point Pair Feature Matching (PPFM). Since its invention in 2010 several modifications have been proposed to cope with its weaknesses, which are computational complexity, sensitivity to noise, and difficulties in the detection of geometrically simple objects with planar surfaces and rotational symmetries. In this work we focus on the latter. We present a novel approach to automatically detect rotational symmetries by matching the object model to itself. Furthermore, we adapt methods for pose verification and use more discriminative features which incorporate global information into the Point Pair Feature. We also examine the effects of other, already existing extensions by testing them on our specialized dataset for geometrically primitive objects. Results show that particularly our handling of symmetries and the augmented features are able to boost recognition rates.",

keywords = "Object detection, Point Pair Features, Pose estimation",

author = "Markus Ziegler and Martin Rudorfer and Xaver Kroischke and Sebastian Krone and J{\"o}rg Kr{\"u}ger",

note = "Publisher Copyright: {\textcopyright} 2019, Springer Nature Switzerland AG.; 12th International Conference on Computer Vision Systems, ICVS 2019 ; Conference date: 23-09-2019 Through 25-09-2019",

year = "2019",

month = nov,

day = "23",

doi = "10.1007/978-3-030-34995-0_40",

language = "English",

isbn = "9783030349943",

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

pages = "445--456",

editor = "Dimitrios Tzovaras and Dimitrios Giakoumis and Markus Vincze and Antonis Argyros",

booktitle = "Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings",

}

Ziegler, M, Rudorfer, M, Kroischke, X, Krone, S & Krüger, J 2019, Point Pair Feature Matching: Evaluating Methods to Detect Simple Shapes. in D Tzovaras, D Giakoumis, M Vincze & A Argyros (eds), Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11754 LNCS, pp. 445-456, 12th International Conference on Computer Vision Systems, ICVS 2019, Thessaloniki, Greece, 23/09/19. https://doi.org/10.1007/978-3-030-34995-0_40

Point Pair Feature Matching: Evaluating Methods to Detect Simple Shapes. / Ziegler, Markus; Rudorfer, Martin; Kroischke, Xaver et al.
Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings. ed. / Dimitrios Tzovaras; Dimitrios Giakoumis; Markus Vincze; Antonis Argyros. 2019. p. 445-456 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11754 LNCS).

Research output: Chapter in Book/Published conference output › Conference publication

TY - GEN

T1 - Point Pair Feature Matching

T2 - 12th International Conference on Computer Vision Systems, ICVS 2019

AU - Ziegler, Markus

AU - Rudorfer, Martin

AU - Kroischke, Xaver

AU - Krone, Sebastian

AU - Krüger, Jörg

PY - 2019/11/23

Y1 - 2019/11/23

N2 - A recent benchmark for 3D object detection and 6D pose estimation from RGB-D images shows the dominance of methods based on Point Pair Feature Matching (PPFM). Since its invention in 2010 several modifications have been proposed to cope with its weaknesses, which are computational complexity, sensitivity to noise, and difficulties in the detection of geometrically simple objects with planar surfaces and rotational symmetries. In this work we focus on the latter. We present a novel approach to automatically detect rotational symmetries by matching the object model to itself. Furthermore, we adapt methods for pose verification and use more discriminative features which incorporate global information into the Point Pair Feature. We also examine the effects of other, already existing extensions by testing them on our specialized dataset for geometrically primitive objects. Results show that particularly our handling of symmetries and the augmented features are able to boost recognition rates.

AB - A recent benchmark for 3D object detection and 6D pose estimation from RGB-D images shows the dominance of methods based on Point Pair Feature Matching (PPFM). Since its invention in 2010 several modifications have been proposed to cope with its weaknesses, which are computational complexity, sensitivity to noise, and difficulties in the detection of geometrically simple objects with planar surfaces and rotational symmetries. In this work we focus on the latter. We present a novel approach to automatically detect rotational symmetries by matching the object model to itself. Furthermore, we adapt methods for pose verification and use more discriminative features which incorporate global information into the Point Pair Feature. We also examine the effects of other, already existing extensions by testing them on our specialized dataset for geometrically primitive objects. Results show that particularly our handling of symmetries and the augmented features are able to boost recognition rates.

KW - Object detection

KW - Point Pair Features

KW - Pose estimation

UR - http://www.scopus.com/inward/record.url?scp=85077015963&partnerID=8YFLogxK

UR - https://link.springer.com/chapter/10.1007/978-3-030-34995-0_40

U2 - 10.1007/978-3-030-34995-0_40

DO - 10.1007/978-3-030-34995-0_40

M3 - Conference publication

AN - SCOPUS:85077015963

SN - 9783030349943

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

SP - 445

EP - 456

BT - Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings

A2 - Tzovaras, Dimitrios

A2 - Giakoumis, Dimitrios

A2 - Vincze, Markus

A2 - Argyros, Antonis

Y2 - 23 September 2019 through 25 September 2019

ER -

Ziegler M, Rudorfer M, Kroischke X, Krone S, Krüger J. Point Pair Feature Matching: Evaluating Methods to Detect Simple Shapes. In Tzovaras D, Giakoumis D, Vincze M, Argyros A, editors, Computer Vision Systems - 12th International Conference, ICVS 2019, Proceedings. 2019. p. 445-456. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-34995-0_40