An adaptable system for RGB-D based human body detection and pose estimation: incorporating attached props

H. Haggag; M. Hossny; S. Nahavandi; O. Haggag

doi:10.1109/SMC.2016.7844458

An adaptable system for RGB-D based human body detection and pose estimation: incorporating attached props

H. Haggag, M. Hossny, S. Nahavandi, O. Haggag

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

16 Citations (Scopus)

Abstract

One of the biggest challenges of RGB-D posture tracking is separating appendages such as briefcases, trolleys, and backpacks from the human body. Markerless motion tracking relies on segmenting each depth frame to a finite set of body parts. This is achieved via supervised learning by assigning each pixel to a certain body part. The training image set for the supervised learning are usually synthesised using popular motion capture databases and an ensemble of 3D models covering a wide range of anthropometric characteristics. In this paper, we propose a novel method for generating training data of human postures with attached objects. The results have shown a significant increase in body-part classification accuracy for subjects with props from 60% to 94% using the generated image set.

Original language	English
Title of host publication	2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings
Editors	Shun-Feng Su
Place of Publication	Piscataway NJ USA
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Pages	1544-1549
Number of pages	6
ISBN (Electronic)	9781509018970, 9781509018192
ISBN (Print)	9781509018987
DOIs	https://doi.org/10.1109/SMC.2016.7844458
Publication status	Published - 2016
Externally published	Yes
Event	IEEE International Conference on Systems, Man and Cybernetics 2016 - Budapest, Hungary Duration: 9 Oct 2016 → 12 Oct 2016 https://ieeexplore.ieee.org/xpl/conhome/7830913/proceeding (Proceedings)

Conference

Conference	IEEE International Conference on Systems, Man and Cybernetics 2016
Abbreviated title	SMC 2016
Country/Territory	Hungary
City	Budapest
Period	9/10/16 → 12/10/16
Internet address	https://ieeexplore.ieee.org/xpl/conhome/7830913/proceeding (Proceedings)

Keywords

Kinect
Pixel Labelling
Props
RGB-D

Access to Document

10.1109/SMC.2016.7844458

Cite this

Haggag, H., Hossny, M., Nahavandi, S., & Haggag, O. (2016). An adaptable system for RGB-D based human body detection and pose estimation: incorporating attached props. In S-F. Su (Ed.), 2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings (pp. 1544-1549). IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/SMC.2016.7844458

Haggag, H. ; Hossny, M. ; Nahavandi, S. et al. / An adaptable system for RGB-D based human body detection and pose estimation : incorporating attached props. 2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings. editor / Shun-Feng Su. Piscataway NJ USA : IEEE, Institute of Electrical and Electronics Engineers, 2016. pp. 1544-1549

@inproceedings{a5e1d8a66af6483e9ae04aaa5a965050,

title = "An adaptable system for RGB-D based human body detection and pose estimation: incorporating attached props",

abstract = "One of the biggest challenges of RGB-D posture tracking is separating appendages such as briefcases, trolleys, and backpacks from the human body. Markerless motion tracking relies on segmenting each depth frame to a finite set of body parts. This is achieved via supervised learning by assigning each pixel to a certain body part. The training image set for the supervised learning are usually synthesised using popular motion capture databases and an ensemble of 3D models covering a wide range of anthropometric characteristics. In this paper, we propose a novel method for generating training data of human postures with attached objects. The results have shown a significant increase in body-part classification accuracy for subjects with props from 60% to 94% using the generated image set.",

keywords = "Kinect, Pixel Labelling, Props, RGB-D",

author = "H. Haggag and M. Hossny and S. Nahavandi and O. Haggag",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.; IEEE International Conference on Systems, Man and Cybernetics 2016, SMC 2016 ; Conference date: 09-10-2016 Through 12-10-2016",

year = "2016",

doi = "10.1109/SMC.2016.7844458",

language = "English",

isbn = "9781509018987",

pages = "1544--1549",

editor = "Shun-Feng Su",

booktitle = "2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

address = "United States of America",

url = "https://ieeexplore.ieee.org/xpl/conhome/7830913/proceeding",

}

Haggag, H, Hossny, M, Nahavandi, S & Haggag, O 2016, An adaptable system for RGB-D based human body detection and pose estimation: incorporating attached props. in S-F Su (ed.), 2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings. IEEE, Institute of Electrical and Electronics Engineers, Piscataway NJ USA, pp. 1544-1549, IEEE International Conference on Systems, Man and Cybernetics 2016, Budapest, Hungary, 9/10/16. https://doi.org/10.1109/SMC.2016.7844458

An adaptable system for RGB-D based human body detection and pose estimation : incorporating attached props. / Haggag, H.; Hossny, M.; Nahavandi, S. et al.

2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings. ed. / Shun-Feng Su. Piscataway NJ USA : IEEE, Institute of Electrical and Electronics Engineers, 2016. p. 1544-1549.

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

TY - GEN

T1 - An adaptable system for RGB-D based human body detection and pose estimation

T2 - IEEE International Conference on Systems, Man and Cybernetics 2016

AU - Haggag, H.

AU - Hossny, M.

AU - Nahavandi, S.

AU - Haggag, O.

PY - 2016

Y1 - 2016

N2 - One of the biggest challenges of RGB-D posture tracking is separating appendages such as briefcases, trolleys, and backpacks from the human body. Markerless motion tracking relies on segmenting each depth frame to a finite set of body parts. This is achieved via supervised learning by assigning each pixel to a certain body part. The training image set for the supervised learning are usually synthesised using popular motion capture databases and an ensemble of 3D models covering a wide range of anthropometric characteristics. In this paper, we propose a novel method for generating training data of human postures with attached objects. The results have shown a significant increase in body-part classification accuracy for subjects with props from 60% to 94% using the generated image set.

AB - One of the biggest challenges of RGB-D posture tracking is separating appendages such as briefcases, trolleys, and backpacks from the human body. Markerless motion tracking relies on segmenting each depth frame to a finite set of body parts. This is achieved via supervised learning by assigning each pixel to a certain body part. The training image set for the supervised learning are usually synthesised using popular motion capture databases and an ensemble of 3D models covering a wide range of anthropometric characteristics. In this paper, we propose a novel method for generating training data of human postures with attached objects. The results have shown a significant increase in body-part classification accuracy for subjects with props from 60% to 94% using the generated image set.

KW - Kinect

KW - Pixel Labelling

KW - Props

KW - RGB-D

UR - http://www.scopus.com/inward/record.url?scp=85015740769&partnerID=8YFLogxK

U2 - 10.1109/SMC.2016.7844458

DO - 10.1109/SMC.2016.7844458

M3 - Conference Paper

AN - SCOPUS:85015740769

SN - 9781509018987

SP - 1544

EP - 1549

BT - 2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings

A2 - Su, Shun-Feng

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Piscataway NJ USA

Y2 - 9 October 2016 through 12 October 2016

ER -

Haggag H, Hossny M, Nahavandi S, Haggag O. An adaptable system for RGB-D based human body detection and pose estimation: incorporating attached props. In Su S-F, editor, 2016 IEEE International Conference on Systems, Man, and Cybernetics - Conference Proceedings. Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers. 2016. p. 1544-1549 https://doi.org/10.1109/SMC.2016.7844458

An adaptable system for RGB-D based human body detection and pose estimation: incorporating attached props

Abstract

Conference

Keywords

Access to Document

Other files and links

Cite this