Towards Efficient Human Action Retrieval based on Triplet-Loss
Metric Learning

D 2022

Towards Efficient Human Action Retrieval based on Triplet-Loss Metric Learning

KICO, Iris, Jan SEDMIDUBSKÝ and Pavel ZEZULA

Basic information

Original name

Towards Efficient Human Action Retrieval based on Triplet-Loss Metric Learning

Authors

KICO, Iris (70 Bosnia and Herzegovina, belonging to the institution), Jan SEDMIDUBSKÝ (203 Czech Republic, guarantor, belonging to the institution) and Pavel ZEZULA (203 Czech Republic, belonging to the institution)

Edition

Berlin, Heidelberg, 33rd International Conference on Database and Expert Systems Applications (DEXA), p. 234-247, 14 pp. 2022

Publisher

Springer-Verlag

Other information

Language

English

Type of outcome

Stať ve sborníku

Field of Study

10200 1.2 Computer and information sciences

Country of publisher

United States of America

Confidentiality degree

není předmětem státního či obchodního tajemství

Publication form

electronic version available online

References:

URL

Impact factor

Impact factor: 0.402 in 2005

RIV identification code

RIV/00216224:14330/22:00125807

Organization unit

Faculty of Informatics

ISBN

978-3-031-12422-8

ISSN

DOI

http://dx.doi.org/10.1007/978-3-031-12423-5_18

UT WoS

000877013800018

Keywords in English

human motion data;skeleton sequences;action similarity;action retrieval;triplet-loss learning;LSTM

Abstract

V originále

Recent pose-estimation methods enable digitization of human motion by extracting 3D skeleton sequences from ordinary video recordings. Such spatio-temporal skeleton representation offers attractive possibilities for a wide range of applications but, at the same time, requires effective and efficient content-based access to make the extracted data reusable. In this paper, we focus on content-based retrieval of pre-segmented skeleton sequences of human actions to identify the most similar ones to a query action. We mainly deal with the extraction of content-preserving action features, which are learned using the triplet-loss approach in an unsupervised way. Such features are (1) effective as they achieve a similar retrieval quality as the features learned in a supervised way, and (2) of a fixed size which enables the application of indexing structures for efficient retrieval.

Links

EF16_019/0000822, research and development project

Name: Centrum excelence pro kyberkriminalitu, kyberbezpečnost a ochranu kritických informačních infrastruktur

MUNI/A/1195/2021, interní kód MU

Name: Aplikovaný výzkum v oblastech vyhledávání, analýz a vizualizací rozsáhlých dat, zpracování přirozeného jazyka a aplikované umělé inteligence

Investor: Masaryk University

Citovat

KICO, Iris, Jan SEDMIDUBSKÝ and Pavel ZEZULA. Towards Efficient Human Action Retrieval based on Triplet-Loss Metric Learning. Online. In Strauss, C., Cuzzocrea, A., Kotsis, G., Tjoa, A.M., Khalil, I. 33rd International Conference on Database and Expert Systems Applications (DEXA). Berlin, Heidelberg: Springer-Verlag, 2022, p. 234-247. ISBN 978-3-031-12422-8. Available from: https://dx.doi.org/10.1007/978-3-031-12423-5_18.

@inproceedings{1852909,
   author = {Kico, Iris and Sedmidubský, Jan and Zezula, Pavel},
   address = {Berlin, Heidelberg},
   booktitle = {33rd International Conference on Database and Expert Systems Applications (DEXA)},
   doi = {http://dx.doi.org/10.1007/978-3-031-12423-5_18},
   editor = {Strauss, C., Cuzzocrea, A., Kotsis, G., Tjoa, A.M., Khalil, I.},
   keywords = {human motion data;skeleton sequences;action similarity;action retrieval;triplet-loss learning;LSTM},
   howpublished = {elektronická verze "online"},
   language = {eng},
   location = {Berlin, Heidelberg},
   isbn = {978-3-031-12422-8},
   pages = {234-247},
   publisher = {Springer-Verlag},
   title = {Towards Efficient Human Action Retrieval based on Triplet-Loss Metric Learning},
   url = {https://link.springer.com/chapter/10.1007/978-3-031-12423-5_18},
   year = {2022}
}

TY  - JOUR
ID  - 1852909
AU  - Kico, Iris - Sedmidubský, Jan - Zezula, Pavel
PY  - 2022
TI  - Towards Efficient Human Action Retrieval based on Triplet-Loss Metric Learning
PB  - Springer-Verlag
CY  - Berlin, Heidelberg
SN  - 9783031124228
KW  - human motion data;skeleton sequences;action similarity;action retrieval;triplet-loss learning;LSTM
UR  - https://link.springer.com/chapter/10.1007/978-3-031-12423-5_18
N2  - Recent pose-estimation methods enable digitization of human motion by extracting 3D skeleton sequences from ordinary video recordings. Such spatio-temporal skeleton representation offers attractive possibilities for a wide range of applications but, at the same time, requires effective and efficient content-based access to make the extracted data reusable. In this paper, we focus on content-based retrieval of pre-segmented skeleton sequences of human actions to identify the most similar ones to a query action. We mainly deal with the extraction of content-preserving action features, which are learned using the triplet-loss approach in an unsupervised way. Such features are (1) effective as they achieve a similar retrieval quality as the features learned in a supervised way, and (2) of a fixed size which enables the application of indexing structures for efficient retrieval.
ER  -

KICO, Iris, Jan SEDMIDUBSKÝ and Pavel ZEZULA. Towards Efficient Human Action Retrieval based on Triplet-Loss Metric Learning. Online. In Strauss, C., Cuzzocrea, A., Kotsis, G., Tjoa, A.M., Khalil, I. \textit{33rd International Conference on Database and Expert Systems Applications (DEXA)}. Berlin, Heidelberg: Springer-Verlag, 2022, p.~234-247. ISBN~978-3-031-12422-8. Available from: https://dx.doi.org/10.1007/978-3-031-12423-5\_{}18.

Detailed Information on Publication Record