Quaternion Watershed Transform in Segmentation of Motion Capture Data

Adam Świtoński; Agnieszka Michalczuk; Henryk Josiński; Konrad Wojciechowski

doi:10.1007/978-3-030-14802-7_49

Quaternion Watershed Transform in Segmentation of Motion Capture Data

Department of Graphics, Computer Vision and Digital Systems (RAU/RAU6)

Polish-Japanese Academy of Information Technology

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

Abstract

The novel approach for segmentation of motion capture data is proposed. It utilizes hierarchical watershed transform for time series containing unit quaternions which describe rotations of human skeleton joints as well as their angular velocities. To approximate gradient magnitudes of subsequent time instants, aggregated geodesic distance on hypersphere (Formula presented) for preceding and following quaternions is computed. The introduced segmentation method was applied to gait analysis. The highly precise motion capture data, registered in a human motion lab (HML), were used. A fully automatic watershed transform with detection of catchment basins as well as a marker controlled one were investigated. The obtained results are promising. By selecting proper hierarchy of a segmentation or by specifying adequate markers, it is even feasible to divide gait cycle into consecutive steps. The segmentation can be improved in respect to a considered problem, if only selected joints are taken into account by watershed transform.

Original language	English
Title of host publication	Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings
Editors	Ford Lumban Gaol, Ngoc Thanh Nguyen, Bogdan Trawiński, Ngoc Thanh Nguyen, Tzung-Pei Hong
Publisher	Springer Verlag
Pages	567-578
Number of pages	12
ISBN (Print)	9783030148010
DOIs	https://doi.org/10.1007/978-3-030-14802-7_49
Publication status	Published - 2019
Event	11th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2019 - Yogyakarta, Indonesia Duration: 8 Apr 2019 → 11 Apr 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11432 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2019
Country/Territory	Indonesia
City	Yogyakarta
Period	8/04/19 → 11/04/19

Keywords

Gait analysis
Motion capture
Motion segmentation
Unit quaternions
Watershed transform

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-030-14802-7_49

Cite this

Świtoński, A., Michalczuk, A., Josiński, H., & Wojciechowski, K. (2019). Quaternion Watershed Transform in Segmentation of Motion Capture Data. In F. L. Gaol, N. T. Nguyen, B. Trawiński, N. T. Nguyen, & T.-P. Hong (Eds.), Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings (pp. 567-578). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11432 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-030-14802-7_49

Świtoński, Adam ; Michalczuk, Agnieszka ; Josiński, Henryk et al. / Quaternion Watershed Transform in Segmentation of Motion Capture Data. Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings. editor / Ford Lumban Gaol ; Ngoc Thanh Nguyen ; Bogdan Trawiński ; Ngoc Thanh Nguyen ; Tzung-Pei Hong. Springer Verlag, 2019. pp. 567-578 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{6b3148ab92624d7ca6b068c4574ebbc9,

title = "Quaternion Watershed Transform in Segmentation of Motion Capture Data",

abstract = "The novel approach for segmentation of motion capture data is proposed. It utilizes hierarchical watershed transform for time series containing unit quaternions which describe rotations of human skeleton joints as well as their angular velocities. To approximate gradient magnitudes of subsequent time instants, aggregated geodesic distance on hypersphere (Formula presented) for preceding and following quaternions is computed. The introduced segmentation method was applied to gait analysis. The highly precise motion capture data, registered in a human motion lab (HML), were used. A fully automatic watershed transform with detection of catchment basins as well as a marker controlled one were investigated. The obtained results are promising. By selecting proper hierarchy of a segmentation or by specifying adequate markers, it is even feasible to divide gait cycle into consecutive steps. The segmentation can be improved in respect to a considered problem, if only selected joints are taken into account by watershed transform.",

keywords = "Gait analysis, Motion capture, Motion segmentation, Unit quaternions, Watershed transform",

author = "Adam {\'S}wito{\'n}ski and Agnieszka Michalczuk and Henryk Josi{\'n}ski and Konrad Wojciechowski",

note = "Publisher Copyright: {\textcopyright} 2019, Springer Nature Switzerland AG.; 11th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2019 ; Conference date: 08-04-2019 Through 11-04-2019",

year = "2019",

doi = "10.1007/978-3-030-14802-7\_49",

language = "English",

isbn = "9783030148010",

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

publisher = "Springer Verlag",

pages = "567--578",

editor = "Gaol, \{Ford Lumban\} and Nguyen, \{Ngoc Thanh\} and Bogdan Trawi{\'n}ski and Nguyen, \{Ngoc Thanh\} and Tzung-Pei Hong",

booktitle = "Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings",

address = "Germany",

}

Świtoński, A , Michalczuk, A , Josiński, H & Wojciechowski, K 2019, Quaternion Watershed Transform in Segmentation of Motion Capture Data. in FL Gaol, NT Nguyen, B Trawiński, NT Nguyen & T-P Hong (eds), Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11432 LNAI, Springer Verlag, pp. 567-578, 11th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2019, Yogyakarta, Indonesia, 8/04/19. https://doi.org/10.1007/978-3-030-14802-7_49

Quaternion Watershed Transform in Segmentation of Motion Capture Data. / Świtoński, Adam ; Michalczuk, Agnieszka ; Josiński, Henryk et al.
Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings. ed. / Ford Lumban Gaol; Ngoc Thanh Nguyen; Bogdan Trawiński; Ngoc Thanh Nguyen; Tzung-Pei Hong. Springer Verlag, 2019. p. 567-578 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11432 LNAI).

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

TY - GEN

T1 - Quaternion Watershed Transform in Segmentation of Motion Capture Data

AU - Świtoński, Adam

AU - Michalczuk, Agnieszka

AU - Josiński, Henryk

AU - Wojciechowski, Konrad

PY - 2019

Y1 - 2019

N2 - The novel approach for segmentation of motion capture data is proposed. It utilizes hierarchical watershed transform for time series containing unit quaternions which describe rotations of human skeleton joints as well as their angular velocities. To approximate gradient magnitudes of subsequent time instants, aggregated geodesic distance on hypersphere (Formula presented) for preceding and following quaternions is computed. The introduced segmentation method was applied to gait analysis. The highly precise motion capture data, registered in a human motion lab (HML), were used. A fully automatic watershed transform with detection of catchment basins as well as a marker controlled one were investigated. The obtained results are promising. By selecting proper hierarchy of a segmentation or by specifying adequate markers, it is even feasible to divide gait cycle into consecutive steps. The segmentation can be improved in respect to a considered problem, if only selected joints are taken into account by watershed transform.

AB - The novel approach for segmentation of motion capture data is proposed. It utilizes hierarchical watershed transform for time series containing unit quaternions which describe rotations of human skeleton joints as well as their angular velocities. To approximate gradient magnitudes of subsequent time instants, aggregated geodesic distance on hypersphere (Formula presented) for preceding and following quaternions is computed. The introduced segmentation method was applied to gait analysis. The highly precise motion capture data, registered in a human motion lab (HML), were used. A fully automatic watershed transform with detection of catchment basins as well as a marker controlled one were investigated. The obtained results are promising. By selecting proper hierarchy of a segmentation or by specifying adequate markers, it is even feasible to divide gait cycle into consecutive steps. The segmentation can be improved in respect to a considered problem, if only selected joints are taken into account by watershed transform.

KW - Gait analysis

KW - Motion capture

KW - Motion segmentation

KW - Unit quaternions

KW - Watershed transform

UR - https://www.scopus.com/pages/publications/85064538197

U2 - 10.1007/978-3-030-14802-7_49

DO - 10.1007/978-3-030-14802-7_49

M3 - Conference contribution

AN - SCOPUS:85064538197

SN - 9783030148010

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

SP - 567

EP - 578

BT - Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings

A2 - Gaol, Ford Lumban

A2 - Nguyen, Ngoc Thanh

A2 - Trawiński, Bogdan

A2 - Nguyen, Ngoc Thanh

A2 - Hong, Tzung-Pei

PB - Springer Verlag

T2 - 11th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2019

Y2 - 8 April 2019 through 11 April 2019

ER -

Świtoński A , Michalczuk A , Josiński H, Wojciechowski K. Quaternion Watershed Transform in Segmentation of Motion Capture Data. In Gaol FL, Nguyen NT, Trawiński B, Nguyen NT, Hong TP, editors, Intelligent Information and Database Systems - 11th Asian Conference, ACIIDS 2019, Proceedings. Springer Verlag. 2019. p. 567-578. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-14802-7_49

Quaternion Watershed Transform in Segmentation of Motion Capture Data

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this