Gesture Recognition Based on Improved HOG-LBP Features

DOI

10.2991/cnci-19.2019.39How to use a DOI?

Keywords

Gesture recognition, HOG-LBP, MB-LBP, SVM.

Abstract

With the development of computer technology, vision-based human gesture recognition has become an important hotspot technology in the field of human-computer interaction. However, the performance of gesture recognition is often affected by conditions such as lighting changes, background complexity, and skin color differences. In this paper, an enhanced fusion HOG feature and LBP feature algorithm are proposed for feature extraction. HOG features describe image local features and LBP features describe image texture features. Improved MB-LBP can capture large-scale structures more than LBP features. It is more able to describe more changes, that is, contain more local information. Then use the Support Vector Machine (SVM) for classification detection. The method is tested in the Jochen Triesch static hand pose data set. The results show that the recognition accuracy reaches 98.64%, which is more robust and effective than single feature extraction and traditional HOG-LBP.

Copyright

© 2019, 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  - Gang Nie
AU  - Junxi Zhao
PY  - 2019/05
DA  - 2019/05
TI  - Gesture Recognition Based on Improved HOG-LBP Features
BT  - Proceedings of the 2019 International Conference on Computer, Network, Communication and Information Systems (CNCI 2019)
PB  - Atlantis Press
SP  - 264
EP  - 268
SN  - 2352-538X
UR  - https://doi.org/10.2991/cnci-19.2019.39
DO  - 10.2991/cnci-19.2019.39
ID  - Nie2019/05
ER  -