Machine Learning and Intelligent Communications. Third International Conference, MLICOM 2018, Hangzhou, China, July 6-8, 2018, Proceedings

Research Article

Physical Violence Detection with Movement Sensors

Download
166 downloads
  • @INPROCEEDINGS{10.1007/978-3-030-00557-3_20,
        author={Liang Ye and Le Wang and Peng Wang and Hany Ferdinando and Tapio Sepp\aa{}nen and Esko Alasaarela},
        title={Physical Violence Detection with Movement Sensors},
        proceedings={Machine Learning and Intelligent Communications. Third International Conference, MLICOM 2018, Hangzhou, China, July 6-8, 2018, Proceedings},
        proceedings_a={MLICOM},
        year={2018},
        month={10},
        keywords={Physical violence detection Activity recognition Movement sensor},
        doi={10.1007/978-3-030-00557-3_20}
    }
    
  • Liang Ye
    Le Wang
    Peng Wang
    Hany Ferdinando
    Tapio Seppänen
    Esko Alasaarela
    Year: 2018
    Physical Violence Detection with Movement Sensors
    MLICOM
    Springer
    DOI: 10.1007/978-3-030-00557-3_20
Liang Ye,*, Le Wang1,*, Peng Wang,*, Hany Ferdinando,*, Tapio Seppänen2,*, Esko Alasaarela2,*
  • 1: Communication Research Center, Harbin Institute of Technology
  • 2: University of Oulu
*Contact email: yeliang@hit.edu.cn, 1659412561@qq.com, wphitstudent@163.com, hferdina@ee.oulu.fi, tapio@ee.oulu.fi, esko.alasaarela@ee.oulu.fi

Abstract

With the development of movement sensors, activity recognition becomes more and more popular. Compared with daily-life activity recognition, physical violence detection is more meaningful and valuable. This paper proposes a physical violence detecting method. Movement data of acceleration and gyro are gathered by role playing of physical violence and daily-life activities. Time domain features and frequency domain ones are extracted and filtered to discribe the differences between physical violence and daily-life activities. A specific BPNN trained with the L-M method works as the classifier. Altogether 9 kinds of activities are involved. For 9-class classification, the average recognition accuracy is 67.0%, whereas for 2-class classification, i.e. activities are classified as violence or daily-life activity, the average recognition accuracy reaches 83.7%.