Physical reasoning for 3D object recognition using global hypothesis verification

Shuichi Akizuki; Manabu Hashimoto

doi:10.1007/978-3-319-49409-8_51

Physical reasoning for 3D object recognition using global hypothesis verification

Shuichi Akizuki, Manabu Hashimoto

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

2 Citations (Scopus)

Abstract

In this paper, we propose a method to recognize the 6DoF pose of multiple objects simultaneously. One good solution to recognize them is applying a Hypothesis Verification (HV) algorithm. This type of algorithm evaluates consistency between an input scene and scene hypotheses represented by combinations of object candidates generated from the model based matching. Its use achieves reliable recognition because it maximizes the fitting score between the input scene and the scene hypotheses instead of maximizing the fitting score of an object candidate. We have developed a more reliable HV algorithm that uses a novel cue, the naturalness of an object’s layout (its physical reasoning). This cue evaluates whether the object’s layout in a scene hypothesis can actually be achieved by using simple collision detection. Experimental results show that using the physical reasoning have improved recognition reliability.

Original language	English
Title of host publication	Computer Vision – ECCV 2016 Workshops, Proceedings
Publisher	Springer Verlag
Pages	595-605
Number of pages	11
Volume	9915 LNCS
ISBN (Print)	9783319494081
DOIs	https://doi.org/10.1007/978-3-319-49409-8_51
Publication status	Published - 2016
Externally published	Yes
Event	14th European Conference on Computer Vision, ECCV 2016 - Amsterdam, Netherlands Duration: 2016 Oct 8 → 2016 Oct 16

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9915 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	14th European Conference on Computer Vision, ECCV 2016
Country	Netherlands
City	Amsterdam
Period	16/10/8 → 16/10/16

Keywords

3D object recognition
Collision detection
Hypothesis verification
Point cloud

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-319-49409-8_51

Fingerprint Dive into the research topics of 'Physical reasoning for 3D object recognition using global hypothesis verification'. Together they form a unique fingerprint.

Cite this

Akizuki, S., & Hashimoto, M. (2016). Physical reasoning for 3D object recognition using global hypothesis verification. In Computer Vision – ECCV 2016 Workshops, Proceedings (Vol. 9915 LNCS, pp. 595-605). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9915 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-49409-8_51

Physical reasoning for 3D object recognition using global hypothesis verification. / Akizuki, Shuichi; Hashimoto, Manabu.

Computer Vision – ECCV 2016 Workshops, Proceedings. Vol. 9915 LNCS Springer Verlag, 2016. p. 595-605 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9915 LNCS).

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

Akizuki, S & Hashimoto, M 2016, Physical reasoning for 3D object recognition using global hypothesis verification. in Computer Vision – ECCV 2016 Workshops, Proceedings. vol. 9915 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9915 LNCS, Springer Verlag, pp. 595-605, 14th European Conference on Computer Vision, ECCV 2016, Amsterdam, Netherlands, 16/10/8. https://doi.org/10.1007/978-3-319-49409-8_51

Akizuki S, Hashimoto M. Physical reasoning for 3D object recognition using global hypothesis verification. In Computer Vision – ECCV 2016 Workshops, Proceedings. Vol. 9915 LNCS. Springer Verlag. 2016. p. 595-605. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-49409-8_51

@inproceedings{c2dac9e003b74a62a6549a98ba16ea9a,

title = "Physical reasoning for 3D object recognition using global hypothesis verification",

abstract = "In this paper, we propose a method to recognize the 6DoF pose of multiple objects simultaneously. One good solution to recognize them is applying a Hypothesis Verification (HV) algorithm. This type of algorithm evaluates consistency between an input scene and scene hypotheses represented by combinations of object candidates generated from the model based matching. Its use achieves reliable recognition because it maximizes the fitting score between the input scene and the scene hypotheses instead of maximizing the fitting score of an object candidate. We have developed a more reliable HV algorithm that uses a novel cue, the naturalness of an object{\textquoteright}s layout (its physical reasoning). This cue evaluates whether the object{\textquoteright}s layout in a scene hypothesis can actually be achieved by using simple collision detection. Experimental results show that using the physical reasoning have improved recognition reliability.",

keywords = "3D object recognition, Collision detection, Hypothesis verification, Point cloud",

author = "Shuichi Akizuki and Manabu Hashimoto",

year = "2016",

doi = "10.1007/978-3-319-49409-8_51",

language = "English",

isbn = "9783319494081",

volume = "9915 LNCS",

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

publisher = "Springer Verlag",

pages = "595--605",

booktitle = "Computer Vision – ECCV 2016 Workshops, Proceedings",

address = "Germany",

note = "14th European Conference on Computer Vision, ECCV 2016 ; Conference date: 08-10-2016 Through 16-10-2016",

}

TY - GEN

T1 - Physical reasoning for 3D object recognition using global hypothesis verification

AU - Akizuki, Shuichi

AU - Hashimoto, Manabu

PY - 2016

Y1 - 2016

N2 - In this paper, we propose a method to recognize the 6DoF pose of multiple objects simultaneously. One good solution to recognize them is applying a Hypothesis Verification (HV) algorithm. This type of algorithm evaluates consistency between an input scene and scene hypotheses represented by combinations of object candidates generated from the model based matching. Its use achieves reliable recognition because it maximizes the fitting score between the input scene and the scene hypotheses instead of maximizing the fitting score of an object candidate. We have developed a more reliable HV algorithm that uses a novel cue, the naturalness of an object’s layout (its physical reasoning). This cue evaluates whether the object’s layout in a scene hypothesis can actually be achieved by using simple collision detection. Experimental results show that using the physical reasoning have improved recognition reliability.

AB - In this paper, we propose a method to recognize the 6DoF pose of multiple objects simultaneously. One good solution to recognize them is applying a Hypothesis Verification (HV) algorithm. This type of algorithm evaluates consistency between an input scene and scene hypotheses represented by combinations of object candidates generated from the model based matching. Its use achieves reliable recognition because it maximizes the fitting score between the input scene and the scene hypotheses instead of maximizing the fitting score of an object candidate. We have developed a more reliable HV algorithm that uses a novel cue, the naturalness of an object’s layout (its physical reasoning). This cue evaluates whether the object’s layout in a scene hypothesis can actually be achieved by using simple collision detection. Experimental results show that using the physical reasoning have improved recognition reliability.

KW - 3D object recognition

KW - Collision detection

KW - Hypothesis verification

KW - Point cloud

UR - http://www.scopus.com/inward/record.url?scp=85006062524&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85006062524&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-49409-8_51

DO - 10.1007/978-3-319-49409-8_51

M3 - Conference contribution

AN - SCOPUS:85006062524

SN - 9783319494081

VL - 9915 LNCS

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

SP - 595

EP - 605

BT - Computer Vision – ECCV 2016 Workshops, Proceedings

PB - Springer Verlag

T2 - 14th European Conference on Computer Vision, ECCV 2016

Y2 - 8 October 2016 through 16 October 2016

ER -