Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

High Power Laser and Particle Beams, Volume. 34, Issue 3, 031023(2022)

Lightweight neural network hand gesture recognition method for embedded platforms

Chenyi Yang, Yuqing He*, Junyuan Zhao, and Guorong Li
Author Affiliations
  • Key Laboratory of Photoelectronic Imaging Technology and System of Ministry of Education, School of Optics and Photonics, Beijing Institute of Technology, Beijing 100081, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (16)
    Equations (4)
    References (15)
    Tools
    Paper Information
    Topics
    Figures & Tables(16)
    Construction and pipeline of the algorithm

    Fig. 1. Construction and pipeline of the algorithm

    Download full size
    View in Article
    Depthwise separable convolution

    Fig. 2. Depthwise separable convolution

    Download full size
    View in Article
    Squeeze-and-excitation module[22]

    Fig. 3. Squeeze-and-excitation module[22]

    Download full size
    View in Article
    Hand gesture recognition network based on MobileNetv3-SSDLite

    Fig. 4. Hand gesture recognition network based on MobileNetv3-SSDLite

    Download full size
    View in Article
    Neural network structure before and after the embedded optimization

    Fig. 5. Neural network structure before and after the embedded optimization

    Download full size
    View in Article
    The chosen hand gestures

    Fig. 6. The chosen hand gestures

    Download full size
    View in Article
    Selecting images for hand gesture dataset

    Fig. 7. Selecting images for hand gesture dataset

    Download full size
    View in Article
    Network loss in training process

    Fig. 8. Network loss in training process

    Download full size
    View in Article
    NVIDIA Jetson TX2 embedded processor developer kit

    Fig. 9. NVIDIA Jetson TX2 embedded processor developer kit

    Download full size
    View in Article
    Part of the hand gesture recognition results

    Fig. 10. Part of the hand gesture recognition results

    Download full size
    View in Article

    • Table 1. MobileNet series comparison to VGG16

      View table
      View in Article

      Table 1. MobileNet series comparison to VGG16

      network structureparams/MbyteMACs/106ImageNet accuracy/%
      VGG1613.81530071.5
      MobieNetv14.256970.6
      MobileNetv23.430072.0
      MobileNetv35.421975.2

    • Table 2. Extra feature map layers for object detection

      View table
      View in Article

      Table 2. Extra feature map layers for object detection

      extra layersshape
      layer 1$39 \times 39 \times 512$
      layer 2$19 \times 19 \times 1024$
      layer 3$10 \times 10 \times 512$
      layer 4$5 \times 5 \times 256$
      layer 5$3 \times 3 \times 256$
      layer 6$1 \times 1 \times 256$

    • Table 3. SSDLite detection head comparison to SSD

      View table
      View in Article

      Table 3. SSDLite detection head comparison to SSD

      network structureparams/MbyteMACs/106mAP/%
      SSD14.8125019.3
      SSDLite2.135022.2

    • Table 4. Recognition results of hand gestures

      View table
      View in Article

      Table 4. Recognition results of hand gestures

      hand gestureaccuracy/%
      099.64
      1100.00
      399.51
      499.22
      599.69
      average99.61

    • Table 5. Recognition results of hand gestureson various scenarios

      View table
      View in Article

      Table 5. Recognition results of hand gestureson various scenarios

      scenariosaverage accuracy/%
      multiple hand gestures96
      complicated background64
      low light intensity72

    • Table 6. Comparison of different hand gesture recognition algorithms.

      View table
      View in Article

      Table 6. Comparison of different hand gesture recognition algorithms.

      algorithmparams/MbyteMACs/106frame rate/(frame/s)mean accuracy/%
      VGG16-SSD24.330654291.75
      MobieNetv1-SSD7.212991293.98
      MobileNetv1-SSDLite4.111301693.86
      MobileNetv2-SSDLite3.16563691.01
      MobileNetv3-SSDLite2.25265899.61
    Tools

    Get Citation

    Copy Citation Text

    Chenyi Yang, Yuqing He, Junyuan Zhao, Guorong Li. Lightweight neural network hand gesture recognition method for embedded platforms[J]. High Power Laser and Particle Beams, 2022, 34(3): 031023

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Category: Laser Advanced Interdisciplinary Science

    Received: Jul. 30, 2021

    Accepted: --

    Published Online: Mar. 28, 2022

    The Author Email: Yuqing He (yuqinghe@bit.edu.cn)

    DOI:10.11884/HPLPB202234.210335

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology