基于Softmax回归分类分析的人体运动检测研究

Research on Human Motion Detection Based on Softmax Regression Classification Analysis

提出一种基于分类分析的人体运动状态识别方法。通过手机内置的加速度传感器采集相关数据,然后对采集的数据进行预处理,采用Softmax回归分类算法对人体运动状态进行分类。在深度学习框架TensorFlow下的实验结果显示此算法分类精度较高,对静止、走路、慢跑、上下楼梯、骑车等五种运动状态的综合识别率为88.18%。

This paper proposes a human motion state recognition method based on classification analysis.The relevant data is collected by the built-in acceleration sensor of the mobile phone,and then the collected data is preprocessed,and the Softmax regression classification algorithm is used to classify the human motion state.The experimental results under the deep learning framework TensorFlow show that the classification accuracy of this algorithm is high,and the comprehensive recognition rate of five kinds of motion states such as static,walking,jogging,up and down stairs,and cycling is 88.18%.