Dynamic gesture classification using skeleton model on RGB-D data

2nd International Conference on Mathematical Modeling in Physical Sciences 2013

物理学;数学

Tamura, Y.^1 ; Umetani, T.^1 ; Kashima, N.^2 ; Nakamura, H.^3

Department of Intelligence and Informatics, Konan University, Okamoto, Kobe, Higashinada, 658-8501, Japan^1

Department of Energy Engineering and Science, Nagoya University, Frocho, Chikusa, 464-8603 Nagoya, Japan^2

Department of Helical Plasma Research, National Institute for Fusion Science, Oroshi, Toki, 509-5292, Japan^3

关键词: Depth camera;    Gesture classifications;    Human gesture recognition;    Natural user interfaces;    Red green blues;    Singular spectrum;    Skeleton models;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/490/1/012103/pdf
DOI  :  10.1088/1742-6596/490/1/012103

来源: IOP

【 摘 要 】

This study aims to subjectively detect and classify similar gestures using a red-green-blue-depth camera. Human gesture recognition is one of the crucial components for realizing natural user interfaces (NUIs) using computers and machines. The quality of the NUI highly depends on the robustness of the achieved gesture recognition. We, therefore, propose a gesture classification method using singular spectrum transformation. Using this method, we can robustly classify gestures and behavior.