A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition

Wei Hong Chin; Kunpei Kato; Azhar Aulia Saputra; Naoyuki Kubota

doi:10.1109/INES46365.2019.9109484

A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition

Wei Hong Chin
, Kunpei Kato
, Azhar Aulia Saputra
, Naoyuki Kubota

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

Abstract

Due to many applications such as video surveillance, human machines interaction and video recovery, human actions recognition was a significant topic in computer vision. This paper proposes a self-organizing recurrent incremental network (SORIN) for human action recognition using human skeleton information. The proposed method models human working memory and episodic memory and comprises two layers of adaptive recurrent Growing-When-Required (ar-GWR) network that connected hierarchically. The working memory layer continually learns incoming perception information and encodes the learned knowledge as neurons. Similarly, the episodic memory layer further learns the spatiotemporal relationship of neurons from working memory as episode neurons to realize human actions incrementally. The proposed method integrates with OpenPose framework for human skeleton action recognition and it is validated through a series of experiments.

Original language	English
Title of host publication	INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	185-190
Number of pages	6
ISBN (Electronic)	9781728112138
DOIs	https://doi.org/10.1109/INES46365.2019.9109484
State	Published - Apr 2019
Externally published	Yes
Event	23rd IEEE International Conference on Intelligent Engineering Systems, INES 2019 - Godollo, Hungary Duration: 25 Apr 2019 → 27 Apr 2019

Publication series

Name	INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings

Conference

Conference	23rd IEEE International Conference on Intelligent Engineering Systems, INES 2019
Country/Territory	Hungary
City	Godollo
Period	25/04/19 → 27/04/19

Bibliographical note

Publisher Copyright:
© 2019 IEEE.

Keywords

Episodic Memory
Human-action Recognition
Incremental Learning
Spatiotemporal Learning

ASJC Scopus subject areas

Artificial Intelligence
Computer Science Applications
Control and Systems Engineering
Mechanical Engineering

Access to Document

10.1109/INES46365.2019.9109484

Cite this

Chin, W. H., Kato, K., Saputra, A. A., & Kubota, N. (2019). A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition. In INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings (pp. 185-190). Article 9109484 (INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/INES46365.2019.9109484

Chin, Wei Hong ; Kato, Kunpei ; Saputra, Azhar Aulia et al. / A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition. INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 185-190 (INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings).

@inproceedings{141bd9abe63d4240858a9d0abbddc585,

title = "A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition",

abstract = "Due to many applications such as video surveillance, human machines interaction and video recovery, human actions recognition was a significant topic in computer vision. This paper proposes a self-organizing recurrent incremental network (SORIN) for human action recognition using human skeleton information. The proposed method models human working memory and episodic memory and comprises two layers of adaptive recurrent Growing-When-Required (ar-GWR) network that connected hierarchically. The working memory layer continually learns incoming perception information and encodes the learned knowledge as neurons. Similarly, the episodic memory layer further learns the spatiotemporal relationship of neurons from working memory as episode neurons to realize human actions incrementally. The proposed method integrates with OpenPose framework for human skeleton action recognition and it is validated through a series of experiments.",

keywords = "Episodic Memory, Human-action Recognition, Incremental Learning, Spatiotemporal Learning",

author = "Chin, \{Wei Hong\} and Kunpei Kato and Saputra, \{Azhar Aulia\} and Naoyuki Kubota",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 23rd IEEE International Conference on Intelligent Engineering Systems, INES 2019 ; Conference date: 25-04-2019 Through 27-04-2019",

year = "2019",

month = apr,

doi = "10.1109/INES46365.2019.9109484",

language = "English",

series = "INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "185--190",

booktitle = "INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings",

address = "United States",

}

Chin, WH, Kato, K, Saputra, AA & Kubota, N 2019, A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition. in INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings., 9109484, INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 185-190, 23rd IEEE International Conference on Intelligent Engineering Systems, INES 2019, Godollo, Hungary, 25/04/19. https://doi.org/10.1109/INES46365.2019.9109484

A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition. / Chin, Wei Hong; Kato, Kunpei; Saputra, Azhar Aulia et al.
INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 185-190 9109484 (INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings).

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

TY - GEN

T1 - A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition

AU - Chin, Wei Hong

AU - Kato, Kunpei

AU - Saputra, Azhar Aulia

AU - Kubota, Naoyuki

PY - 2019/4

Y1 - 2019/4

N2 - Due to many applications such as video surveillance, human machines interaction and video recovery, human actions recognition was a significant topic in computer vision. This paper proposes a self-organizing recurrent incremental network (SORIN) for human action recognition using human skeleton information. The proposed method models human working memory and episodic memory and comprises two layers of adaptive recurrent Growing-When-Required (ar-GWR) network that connected hierarchically. The working memory layer continually learns incoming perception information and encodes the learned knowledge as neurons. Similarly, the episodic memory layer further learns the spatiotemporal relationship of neurons from working memory as episode neurons to realize human actions incrementally. The proposed method integrates with OpenPose framework for human skeleton action recognition and it is validated through a series of experiments.

AB - Due to many applications such as video surveillance, human machines interaction and video recovery, human actions recognition was a significant topic in computer vision. This paper proposes a self-organizing recurrent incremental network (SORIN) for human action recognition using human skeleton information. The proposed method models human working memory and episodic memory and comprises two layers of adaptive recurrent Growing-When-Required (ar-GWR) network that connected hierarchically. The working memory layer continually learns incoming perception information and encodes the learned knowledge as neurons. Similarly, the episodic memory layer further learns the spatiotemporal relationship of neurons from working memory as episode neurons to realize human actions incrementally. The proposed method integrates with OpenPose framework for human skeleton action recognition and it is validated through a series of experiments.

KW - Episodic Memory

KW - Human-action Recognition

KW - Incremental Learning

KW - Spatiotemporal Learning

UR - https://www.scopus.com/pages/publications/85086726869

U2 - 10.1109/INES46365.2019.9109484

DO - 10.1109/INES46365.2019.9109484

M3 - Conference contribution

AN - SCOPUS:85086726869

T3 - INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings

SP - 185

EP - 190

BT - INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 23rd IEEE International Conference on Intelligent Engineering Systems, INES 2019

Y2 - 25 April 2019 through 27 April 2019

ER -

Chin WH, Kato K, Saputra AA, Kubota N. A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition. In INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. p. 185-190. 9109484. (INES 2019 - IEEE 23rd International Conference on Intelligent Engineering Systems, Proceedings). doi: 10.1109/INES46365.2019.9109484

A Hierarchical Dual-Memory Learning Model for Human Skeleton Action Recognition

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this