This paper proposes a multi-level feature learning framework for human action recognition using a single body-worn inertial sensor. The framework consists of three phases, respectively designed to analyze signal-based (low-level), components (mid-level) and semantic (high-level) information. Low-level features capture the time and frequency domain property while mid-level representations learn the composition of the action. The Max-margin Latent Pattern Learning (MLPL) method is proposed to learn high-level semantic descriptions of latent action patterns as the output of our framework. The proposed method achieves the state-of-the-art performances, 88.7%, 98.8% and 72.6% (weighted F1 score) respectively, on Skoda, WISDM and OPP datasets.
@article{arxiv.1611.07143,
title = {Learning Multi-level Features For Sensor-based Human Action Recognition},
author = {Yan Xu and Zhengyang Shen and Xin Zhang and Yifan Gao and Shujian Deng and Yipei Wang and Yubo Fan and Eric I-Chao Chang},
journal= {arXiv preprint arXiv:1611.07143},
year = {2017}
}