Action Recognition Based on Machine Learning under Multiple Motion Sensing Cameras

Project Details


Recognition and analysis of human action are one of the most popular issues in the computer vision and pattern recognition community.Human action consists of consecutive static poses and contains complicated highdimensional information in spatial-temporal domain. There still exist bottlenecks in precise analysis of human action by single view RGB-D camera due to technical problems, like the self-shadowing phenomenon of human poses.Intuitively, human skeleton is appropriate representation of human profile. So we propose the scenarios using multi RGB-D cameras to resolve the self-shadowing problem and adopting the human skeleton axis for object tracking and action recognition.In the proposed system, human skeleton detection of open source is adopted for the basis of action recognition and we can get a large amount of metadata. Machine learning technique is used to classify these metadata to build our own dataset.We expect that the effort will have positive impact on various applications like anti-theft security, home care, physical rehabilitation, and traffic accident prevention. The research results can not only facilitate the relevant studies but also motivate commercial opportunities.
Effective start/end date1/08/1631/07/17

UN Sustainable Development Goals

In 2015, UN member states agreed to 17 global Sustainable Development Goals (SDGs) to end poverty, protect the planet and ensure prosperity for all. This project contributes towards the following SDG(s):

  • SDG 8 - Decent Work and Economic Growth
  • SDG 11 - Sustainable Cities and Communities
  • SDG 17 - Partnerships for the Goals


  • machine learning
  • action recognition
  • RGB-D camera


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.