Motion Recognition Employing Multiple Kernel Learning of Fisher Vectors Using Local Skeleton Features
Abstract
We propose a skeleton-based motion recognition system focusing on local parts of the human body closely related to a target motion. In this system, a skeleton feature is composed of a sequence of relative positions between paired joints calculated by Inverse Kinematics. Several joints of skeleton model are connected as a Local Skeleton Feature. The temporal sequence is modeled as human motion model by using Hidden Markov Model. Motion features are represented as Fisher vectors parameterized by the human motion models, and weighted and integrated by using Multiple Kernel Learning. This system makes it possible for robots to recognize human actions in our daily life. The experimental results based on two datasets show an improvement in performance of classification rate, which shows that the design of motion feature is effective for motion recognition.
Cite
Text
Goutsu et al. "Motion Recognition Employing Multiple Kernel Learning of Fisher Vectors Using Local Skeleton Features." IEEE/CVF International Conference on Computer Vision Workshops, 2015. doi:10.1109/ICCVW.2015.50Markdown
[Goutsu et al. "Motion Recognition Employing Multiple Kernel Learning of Fisher Vectors Using Local Skeleton Features." IEEE/CVF International Conference on Computer Vision Workshops, 2015.](https://mlanthology.org/iccvw/2015/goutsu2015iccvw-motion/) doi:10.1109/ICCVW.2015.50BibTeX
@inproceedings{goutsu2015iccvw-motion,
title = {{Motion Recognition Employing Multiple Kernel Learning of Fisher Vectors Using Local Skeleton Features}},
author = {Goutsu, Yusuke and Takano, Wataru and Nakamura, Yoshihiko},
booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
year = {2015},
pages = {321-328},
doi = {10.1109/ICCVW.2015.50},
url = {https://mlanthology.org/iccvw/2015/goutsu2015iccvw-motion/}
}