Toward a Real Time View-invariant 3D Action Recognition

Mounir Hammouche, Enjie Ghorbel, Anthony Fleury, Sébastien Ambellouis

2016

Abstract

In this paper we propose a novel human action recognition method, robust to viewpoint variation, which combines skeleton- and depth-based action recognition approaches. For this matter, we first build several base classifiers, to independently predict the action performed by a subject. Then, two efficient combination strategies, that take into account skeleton accuracy and human body orientation, are proposed. The first is based on fuzzy switcher where the second uses a combination between fuzzy switcher and aggregation. Moreover, we introduce a new algorithm for the estimation of human body orientation. To perform the test we have created a new Multiview 3D Action public dataset with three viewpoint angles (30°,0°,-30°). The experimental results show that an efficient combination strategy of base classifiers improves the accuracy and the computational efficiency for human action recognition.

Paper Citation

in Harvard Style

Hammouche M., Ghorbel E., Fleury A. and Ambellouis S. (2016). Toward a Real Time View-invariant 3D Action Recognition . In Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISION4HCI, (VISIGRAPP 2016) ISBN 978-989-758-175-5, pages 745-754. DOI: 10.5220/0005843607450754

in Bibtex Style

@conference{vision4hci16,
author={Mounir Hammouche and Enjie Ghorbel and Anthony Fleury and Sébastien Ambellouis},
title={Toward a Real Time View-invariant 3D Action Recognition},
booktitle={Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISION4HCI, (VISIGRAPP 2016)},
year={2016},
pages={745-754},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005843607450754},
isbn={978-989-758-175-5},
}

in EndNote Style

TY - CONF
JO - Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISION4HCI, (VISIGRAPP 2016)
TI - Toward a Real Time View-invariant 3D Action Recognition
SN - 978-989-758-175-5
AU - Hammouche M.
AU - Ghorbel E.
AU - Fleury A.
AU - Ambellouis S.
PY - 2016
SP - 745
EP - 754
DO - 10.5220/0005843607450754