3D Hand Trajectory Recognition with H-ELM

DOI

10.2991/ncce-18.2018.63How to use a DOI?

Keywords

gesture recognition; 3D trajectory; U-chord curvature; H-ELM.

Abstract

In this paper, we present a method to extract features of dynamic gestures and use the Hierarchical Extreme Learning Machine (H-ELM) for gesture recognition. We use a Kinect sensor to record the motion of the three joint points of palm, wrist and elbow. The relation among the three trajectories is extracted as a gesture feature. Then the key nodes in the trajectory are extracted by calculating the U-Chord Curvature algorithm for eliminating redundant nodes and simplifying calculation. Then, through the sparse automatic coding and hierarchical training of the H-ELM, the input of automatic coding is approximate to the original input, and the reconstruction error is reduced. The experiment proves that H-ELM is faster than SVM and original ELM, and the recognition accuracy is higher

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Jingjing Gao
AU  - Yinwei Zhan
PY  - 2018/05
DA  - 2018/05
TI  - 3D Hand Trajectory Recognition with H-ELM
BT  - Proceedings of the 2018 International Conference on Network, Communication, Computer Engineering (NCCE 2018)
PB  - Atlantis Press
SP  - 389
EP  - 395
SN  - 1951-6851
UR  - https://doi.org/10.2991/ncce-18.2018.63
DO  - 10.2991/ncce-18.2018.63
ID  - Gao2018/05
ER  -