Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette

Takuya Ushinohama, Yosuke Sawai, Satoshi Ono, Hiroshi Kawasaki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper proposes a method for simultaneous global registration of multiple depth images which are obtained from multiple viewpoints. Unlike the previous method, the proposed method fully utilizes a silhouette-based cost function taking out-of-view and non-overlapping regions into account as well as depth differences at overlapping areas. With the combination of the above cost functions and a recent powerful meta-heuristics named self-adaptive Differential Evolution, it realizes the entire shape reconstruction from relatively small number (three or four) of depth images, which do not involve enough overlapping regions for Iterative Closest Point even if they are prealigned. In addition, to allow the technique to be applicable not only to time-of-flight sensors, but also projector-camera systems, which has deficient silhouette by occlusions, we propose a simple solution based on color-based silhouette. Experimental results show that the proposed method can reconstruct the entire shape only from three depth images of both synthetic and real data. The influence of noises and inaccurate silhouettes is also evaluated.

Original languageEnglish
Title of host publicationComputer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers
EditorsMing-Hsuan Yang, Hideo Saito, Daniel Cremers, Ian Reid
PublisherSpringer Verlag
Pages457-472
Number of pages16
ISBN (Print)9783319168074
DOIs
Publication statusPublished - Jan 1 2015
Event12th Asian Conference on Computer Vision, ACCV 2014 - Singapore, Singapore
Duration: Nov 1 2014Nov 5 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9004
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other12th Asian Conference on Computer Vision, ACCV 2014
CountrySingapore
CitySingapore
Period11/1/1411/5/14

Fingerprint

Silhouette
Cost functions
Registration
Entire
Overlapping
Cost Function
Cameras
Color
Shape Reconstruction
Time-of-flight
Sensors
Projector
Differential Evolution
Inaccurate
Occlusion
Metaheuristics
Camera
Sensor
Experimental Results

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Ushinohama, T., Sawai, Y., Ono, S., & Kawasaki, H. (2015). Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette. In M-H. Yang, H. Saito, D. Cremers, & I. Reid (Eds.), Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers (pp. 457-472). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9004). Springer Verlag. https://doi.org/10.1007/978-3-319-16808-1_31

Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette. / Ushinohama, Takuya; Sawai, Yosuke; Ono, Satoshi; Kawasaki, Hiroshi.

Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers. ed. / Ming-Hsuan Yang; Hideo Saito; Daniel Cremers; Ian Reid. Springer Verlag, 2015. p. 457-472 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9004).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ushinohama, T, Sawai, Y, Ono, S & Kawasaki, H 2015, Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette. in M-H Yang, H Saito, D Cremers & I Reid (eds), Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9004, Springer Verlag, pp. 457-472, 12th Asian Conference on Computer Vision, ACCV 2014, Singapore, Singapore, 11/1/14. https://doi.org/10.1007/978-3-319-16808-1_31
Ushinohama T, Sawai Y, Ono S, Kawasaki H. Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette. In Yang M-H, Saito H, Cremers D, Reid I, editors, Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers. Springer Verlag. 2015. p. 457-472. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-16808-1_31
Ushinohama, Takuya ; Sawai, Yosuke ; Ono, Satoshi ; Kawasaki, Hiroshi. / Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette. Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers. editor / Ming-Hsuan Yang ; Hideo Saito ; Daniel Cremers ; Ian Reid. Springer Verlag, 2015. pp. 457-472 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{90cfce460aff44d0953e652c4e05fe60,
title = "Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette",
abstract = "This paper proposes a method for simultaneous global registration of multiple depth images which are obtained from multiple viewpoints. Unlike the previous method, the proposed method fully utilizes a silhouette-based cost function taking out-of-view and non-overlapping regions into account as well as depth differences at overlapping areas. With the combination of the above cost functions and a recent powerful meta-heuristics named self-adaptive Differential Evolution, it realizes the entire shape reconstruction from relatively small number (three or four) of depth images, which do not involve enough overlapping regions for Iterative Closest Point even if they are prealigned. In addition, to allow the technique to be applicable not only to time-of-flight sensors, but also projector-camera systems, which has deficient silhouette by occlusions, we propose a simple solution based on color-based silhouette. Experimental results show that the proposed method can reconstruct the entire shape only from three depth images of both synthetic and real data. The influence of noises and inaccurate silhouettes is also evaluated.",
author = "Takuya Ushinohama and Yosuke Sawai and Satoshi Ono and Hiroshi Kawasaki",
year = "2015",
month = "1",
day = "1",
doi = "10.1007/978-3-319-16808-1_31",
language = "English",
isbn = "9783319168074",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "457--472",
editor = "Ming-Hsuan Yang and Hideo Saito and Daniel Cremers and Ian Reid",
booktitle = "Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers",
address = "Germany",

}

TY - GEN

T1 - Simultaneous entire shape registration of multiple depth images using depth difference and shape silhouette

AU - Ushinohama, Takuya

AU - Sawai, Yosuke

AU - Ono, Satoshi

AU - Kawasaki, Hiroshi

PY - 2015/1/1

Y1 - 2015/1/1

N2 - This paper proposes a method for simultaneous global registration of multiple depth images which are obtained from multiple viewpoints. Unlike the previous method, the proposed method fully utilizes a silhouette-based cost function taking out-of-view and non-overlapping regions into account as well as depth differences at overlapping areas. With the combination of the above cost functions and a recent powerful meta-heuristics named self-adaptive Differential Evolution, it realizes the entire shape reconstruction from relatively small number (three or four) of depth images, which do not involve enough overlapping regions for Iterative Closest Point even if they are prealigned. In addition, to allow the technique to be applicable not only to time-of-flight sensors, but also projector-camera systems, which has deficient silhouette by occlusions, we propose a simple solution based on color-based silhouette. Experimental results show that the proposed method can reconstruct the entire shape only from three depth images of both synthetic and real data. The influence of noises and inaccurate silhouettes is also evaluated.

AB - This paper proposes a method for simultaneous global registration of multiple depth images which are obtained from multiple viewpoints. Unlike the previous method, the proposed method fully utilizes a silhouette-based cost function taking out-of-view and non-overlapping regions into account as well as depth differences at overlapping areas. With the combination of the above cost functions and a recent powerful meta-heuristics named self-adaptive Differential Evolution, it realizes the entire shape reconstruction from relatively small number (three or four) of depth images, which do not involve enough overlapping regions for Iterative Closest Point even if they are prealigned. In addition, to allow the technique to be applicable not only to time-of-flight sensors, but also projector-camera systems, which has deficient silhouette by occlusions, we propose a simple solution based on color-based silhouette. Experimental results show that the proposed method can reconstruct the entire shape only from three depth images of both synthetic and real data. The influence of noises and inaccurate silhouettes is also evaluated.

UR - http://www.scopus.com/inward/record.url?scp=84945945341&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84945945341&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-16808-1_31

DO - 10.1007/978-3-319-16808-1_31

M3 - Conference contribution

SN - 9783319168074

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

SP - 457

EP - 472

BT - Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers

A2 - Yang, Ming-Hsuan

A2 - Saito, Hideo

A2 - Cremers, Daniel

A2 - Reid, Ian

PB - Springer Verlag

ER -