基于行为识别匹配的用户日常运动检测

Detection of User

摘要:

针对用户日常运动过程中步态计步器(Pedometer of Gait, POG)检测问题，相较于传统的统计机器学习方法，基于模板匹配的方法往往具有更高的精度和效率。主要提出了一种基于模板匹配的改进Dynamic Template Wrapping(DTW)方法，并在开源实验数据上做了验证。实验结果显示，相对于传统模板匹配方法和统计学习方法，所提出的IsDTW方法不仅具有相对较高的实验精度(92%)，在运行效率上也在一定程度上优于传统方法，对实际应用更具价值。

Abstract:

According to the user's daily movements during gait (Pedometer of Gait POG pedometer) detection, statistical machine learning methods compared with the traditional method, based on template matching has higher precision and efficiency. This paper proposes an improved Dynamic Template Wrapping (DTW) method based on template matching, which is verified on open source experimental data. The experimental results show that compared with the traditional template matching method and IsDTW statistical learning method, the proposed method not only has the relatively high accuracy (92%), the efficiency is better than the traditional method in a certain extent, to the practical application of more value.