Gradient local auto-correlations and extreme learning machine for depth-based activity recognition

Chen Chen; Zhenjie Hou; Baochang Zhang; Junjun Jiang; Yun Yang

doi:10.1007/978-3-319-27857-5_55

Gradient local auto-correlations and extreme learning machine for depth-based activity recognition

Chen Chen
, Zhenjie Hou^*
, Baochang Zhang
, Junjun Jiang
, Yun Yang

^*Corresponding author for this work

School of Artificial Intelligence(Institute of Artificial Intelligence)

University of Texas at Dallas
Changzhou University
China University of Geosciences, Wuhan
Beihang University

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

27 Scopus citations

Abstract

This paper presents a new method for human activity recognition using depth sequences. Each depth sequence is represented by three depth motion maps (DMMs) from three projection views (front, side and top) to capture motion cues. A feature extraction method utilizing spatial and orientational auto-correlations of image local gradients is introduced to extract features from DMMs. The gradient local auto-correlations (GLAC) method employs second order statistics (i.e., auto-correlations) to capture richer information from images than the histogram-based methods (e.g., histogram of oriented gradients) which use first order statistics (i.e., histograms). Based on the extreme learning machine, a fusion framework that incorporates feature-level fusion into decision-level fusion is proposed to effectively combine the GLAC features from DMMs. Experiments on the MSRAction3D and MSRGesture3D datasets demonstrate the effectiveness of the proposed activity recognition algorithm.

Original language	English
Title of host publication	Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings
Editors	Mark Elendt, Richard Boyle, Eric Ragan, Bahram Parvin, Rogerio Feris, Tim McGraw, Ioannis Pavlidis, Regis Kopper, George Bebis, Darko Koracin, Zhao Ye, Gunther Weber
Publisher	Springer Verlag
Pages	613-623
Number of pages	11
ISBN (Print)	9783319278568
DOIs	https://doi.org/10.1007/978-3-319-27857-5_55
State	Published - 2015
Event	11th International Symposium on Advances in Visual Computing, ISVC 2015 - Las Vegas, United States Duration: 14 Dec 2015 → 16 Dec 2015

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9474
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th International Symposium on Advances in Visual Computing, ISVC 2015
Country/Territory	United States
City	Las Vegas
Period	14/12/15 → 16/12/15

Keywords

Activity recognition
Depth images
Depth motion map
Extreme learning machine
Gradient local auto-correlations

Access to Document

10.1007/978-3-319-27857-5_55

Cite this

Chen, C., Hou, Z., Zhang, B., Jiang, J., & Yang, Y. (2015). Gradient local auto-correlations and extreme learning machine for depth-based activity recognition. In M. Elendt, R. Boyle, E. Ragan, B. Parvin, R. Feris, T. McGraw, I. Pavlidis, R. Kopper, G. Bebis, D. Koracin, Z. Ye, & G. Weber (Eds.), Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings (pp. 613-623). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9474). Springer Verlag. https://doi.org/10.1007/978-3-319-27857-5_55

Chen, Chen ; Hou, Zhenjie ; Zhang, Baochang et al. / Gradient local auto-correlations and extreme learning machine for depth-based activity recognition. Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings. editor / Mark Elendt ; Richard Boyle ; Eric Ragan ; Bahram Parvin ; Rogerio Feris ; Tim McGraw ; Ioannis Pavlidis ; Regis Kopper ; George Bebis ; Darko Koracin ; Zhao Ye ; Gunther Weber. Springer Verlag, 2015. pp. 613-623 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{c734f13ab2714f5597e34db3d13aae23,

title = "Gradient local auto-correlations and extreme learning machine for depth-based activity recognition",

abstract = "This paper presents a new method for human activity recognition using depth sequences. Each depth sequence is represented by three depth motion maps (DMMs) from three projection views (front, side and top) to capture motion cues. A feature extraction method utilizing spatial and orientational auto-correlations of image local gradients is introduced to extract features from DMMs. The gradient local auto-correlations (GLAC) method employs second order statistics (i.e., auto-correlations) to capture richer information from images than the histogram-based methods (e.g., histogram of oriented gradients) which use first order statistics (i.e., histograms). Based on the extreme learning machine, a fusion framework that incorporates feature-level fusion into decision-level fusion is proposed to effectively combine the GLAC features from DMMs. Experiments on the MSRAction3D and MSRGesture3D datasets demonstrate the effectiveness of the proposed activity recognition algorithm.",

keywords = "Activity recognition, Depth images, Depth motion map, Extreme learning machine, Gradient local auto-correlations",

author = "Chen Chen and Zhenjie Hou and Baochang Zhang and Junjun Jiang and Yun Yang",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 11th International Symposium on Advances in Visual Computing, ISVC 2015 ; Conference date: 14-12-2015 Through 16-12-2015",

year = "2015",

doi = "10.1007/978-3-319-27857-5\_55",

language = "英语",

isbn = "9783319278568",

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

publisher = "Springer Verlag",

pages = "613--623",

editor = "Mark Elendt and Richard Boyle and Eric Ragan and Bahram Parvin and Rogerio Feris and Tim McGraw and Ioannis Pavlidis and Regis Kopper and George Bebis and Darko Koracin and Zhao Ye and Gunther Weber",

booktitle = "Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings",

address = "德国",

}

Chen, C, Hou, Z, Zhang, B, Jiang, J & Yang, Y 2015, Gradient local auto-correlations and extreme learning machine for depth-based activity recognition. in M Elendt, R Boyle, E Ragan, B Parvin, R Feris, T McGraw, I Pavlidis, R Kopper, G Bebis, D Koracin, Z Ye & G Weber (eds), Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9474, Springer Verlag, pp. 613-623, 11th International Symposium on Advances in Visual Computing, ISVC 2015, Las Vegas, United States, 14/12/15. https://doi.org/10.1007/978-3-319-27857-5_55

Gradient local auto-correlations and extreme learning machine for depth-based activity recognition. / Chen, Chen; Hou, Zhenjie; Zhang, Baochang et al.
Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings. ed. / Mark Elendt; Richard Boyle; Eric Ragan; Bahram Parvin; Rogerio Feris; Tim McGraw; Ioannis Pavlidis; Regis Kopper; George Bebis; Darko Koracin; Zhao Ye; Gunther Weber. Springer Verlag, 2015. p. 613-623 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9474).

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

TY - GEN

T1 - Gradient local auto-correlations and extreme learning machine for depth-based activity recognition

AU - Chen, Chen

AU - Hou, Zhenjie

AU - Zhang, Baochang

AU - Jiang, Junjun

AU - Yang, Yun

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - This paper presents a new method for human activity recognition using depth sequences. Each depth sequence is represented by three depth motion maps (DMMs) from three projection views (front, side and top) to capture motion cues. A feature extraction method utilizing spatial and orientational auto-correlations of image local gradients is introduced to extract features from DMMs. The gradient local auto-correlations (GLAC) method employs second order statistics (i.e., auto-correlations) to capture richer information from images than the histogram-based methods (e.g., histogram of oriented gradients) which use first order statistics (i.e., histograms). Based on the extreme learning machine, a fusion framework that incorporates feature-level fusion into decision-level fusion is proposed to effectively combine the GLAC features from DMMs. Experiments on the MSRAction3D and MSRGesture3D datasets demonstrate the effectiveness of the proposed activity recognition algorithm.

AB - This paper presents a new method for human activity recognition using depth sequences. Each depth sequence is represented by three depth motion maps (DMMs) from three projection views (front, side and top) to capture motion cues. A feature extraction method utilizing spatial and orientational auto-correlations of image local gradients is introduced to extract features from DMMs. The gradient local auto-correlations (GLAC) method employs second order statistics (i.e., auto-correlations) to capture richer information from images than the histogram-based methods (e.g., histogram of oriented gradients) which use first order statistics (i.e., histograms). Based on the extreme learning machine, a fusion framework that incorporates feature-level fusion into decision-level fusion is proposed to effectively combine the GLAC features from DMMs. Experiments on the MSRAction3D and MSRGesture3D datasets demonstrate the effectiveness of the proposed activity recognition algorithm.

KW - Activity recognition

KW - Depth images

KW - Depth motion map

KW - Extreme learning machine

KW - Gradient local auto-correlations

UR - https://www.scopus.com/pages/publications/84952663842

U2 - 10.1007/978-3-319-27857-5_55

DO - 10.1007/978-3-319-27857-5_55

M3 - 会议稿件

AN - SCOPUS:84952663842

SN - 9783319278568

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

SP - 613

EP - 623

BT - Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings

A2 - Elendt, Mark

A2 - Boyle, Richard

A2 - Ragan, Eric

A2 - Parvin, Bahram

A2 - Feris, Rogerio

A2 - McGraw, Tim

A2 - Pavlidis, Ioannis

A2 - Kopper, Regis

A2 - Bebis, George

A2 - Koracin, Darko

A2 - Ye, Zhao

A2 - Weber, Gunther

PB - Springer Verlag

T2 - 11th International Symposium on Advances in Visual Computing, ISVC 2015

Y2 - 14 December 2015 through 16 December 2015

ER -

Chen C, Hou Z, Zhang B, Jiang J, Yang Y. Gradient local auto-correlations and extreme learning machine for depth-based activity recognition. In Elendt M, Boyle R, Ragan E, Parvin B, Feris R, McGraw T, Pavlidis I, Kopper R, Bebis G, Koracin D, Ye Z, Weber G, editors, Advances in Visual Computing - 11th International Symposium, ISVC 2015, Proceedings. Springer Verlag. 2015. p. 613-623. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-27857-5_55

Gradient local auto-correlations and extreme learning machine for depth-based activity recognition

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this