Fall Detection Algorithm Based on Lightweight Openpose Model with Attention Mechanism


  • Ruiming Qiu
  • Wei Teng
  • Zihe Wei
  • Cong Zhang
  • Yipeng Zhong
  • Junhao Zhai




Fall detection, Lightweight Openpose, Attention mechanism.


At present, the serious consequences caused by the fall of the elderly emerge one after another, and installing cameras at home to prevent emergencies for the elderly has gradually become a choice for more and more people. However, the traditional Openpose model cannot detect in real time, which is inconsistent with the actual demand. Therefore, this paper first proposes to use mobilenetv2 lightweight network to replace the original huge vgg19 backbone network, which makes the model real-time in actual use, and then integrates the attention mechanism module to improve the accuracy of the model without affecting the real-time.


Liu Xiaohong, Wu Miao, Niu Qian:Risk Factors of Falls in the Elderly, Beijing Med, Vol. 43 (2021) No.6,p.533-534+538.

Zhang Zezheng, Wang Jun, Dong Mingli, Wang Lei, Yan Bixi: Rapid Key Point Detection Method for Humanoid Robot Based on Improved OpenPose, Laser Journal, [2022-10-18] ,p.1-7.

Yang Ye, Jie Qiang Zhang: Research on Maize Disease Recognition based on Lightweight network MobileNetV2, Modern Computer, Vol. 28 (2022) No.11,p.46-50.

Ji Guangkai, Wang Rong, Peng Shufan: Pedestrian re-recognition method based on attention mechanism and conditional convolution, Journal of Beijing University of Aeronautics and Astronautics , [2022-10-18],p.1-10.



13 November 2022

How to Cite

Qiu, R., Teng, W., Wei, Z., Zhang, C., Zhong, Y., & Zhai, J. (2022). Fall Detection Algorithm Based on Lightweight Openpose Model with Attention Mechanism. Academic Journal of Science and Technology, 3(3), 12–14. https://doi.org/10.54097/ajst.v3i3.2525