Laser & Optoelectronics Progress, Volume. 57, Issue 10, 101008(2020)

Expression Recognition Based on Low Pixel Face Images

Fu Liu, Maojun Li*, Jianwen Hu, Yuhe Xiao, and Zhan Qi
Author Affiliations
  • School of Electrical and Information Engineering, Changsha University of Science and Technology, Changsha, Hunan 410114, China
  • show less
    Figures & Tables(9)
    Pipeline of the facial expression preprocessing
    CNN model
    8 kinds of facial expression images
    Diagram of SoftMax average voting
    • Table 1. Convolutional network parameters

      View table

      Table 1. Convolutional network parameters

      LayerTypeKernel size /pixel×pixelStrideOutput numberOutput size /pixel×pixel
      Layer 1Convolution3×316430×30
      Layer 2Convolution3×3112828×28
      Layer 3Max pooling2×2212814×14
      Layer 4Convolution3×3112814×14
      Layer 5Convolution3×3112814×14
      Layer 6Convolution3×3125612×12
      Layer 7Max pooling2×222566×6
      Layer 8Convolution3×312566×6
      Layer 9Convolution3×312566×6
      Layer 10Average pooling3×332562×2
      Layer 11Convolution2×215121×1
      Layer 12Fully connected---160×1
      OutputSoftMax---8×1
    • Table 2. Facial expression dataset

      View table

      Table 2. Facial expression dataset

      CategoryAngryDisgustFearfulHappySadSurprisedScornNeutral
      Train database14040122207800143008970858054608970
      Test database2724152815161215
    • Table 3. Comparison of recognition accuracy and time

      View table

      Table 3. Comparison of recognition accuracy and time

      ModelOriginaldatabaseLocal binarypatterndatabaseHistogramequalizationdatabase
      Accuracy /%89.586.990.5
      Time/s1.541.271.56
    • Table 4. Improved recognition accuracy of CNN modelunit: %

      View table

      Table 4. Improved recognition accuracy of CNN modelunit: %

      Category AccuracyAngryDisgustFearfulHappySadSurprisedScornNeutralTotal
      First experiment96.391.760.0100.086.7100.083.393.390.5
      Second experiment88.991.786.7100.080.0100.083.393.391.2
      Third experiment96.395.860.0100.086.793.875.093.389.8
      Total93.893.168.9100.084.597.980.593.390.5
    • Table 5. Comparison of recognition accuracy and time between two models

      View table

      Table 5. Comparison of recognition accuracy and time between two models

      ModelLeNet-5Without decision level
      Accuracy /%74.687.9
      Time /s0.590.31
    Tools

    Get Citation

    Copy Citation Text

    Fu Liu, Maojun Li, Jianwen Hu, Yuhe Xiao, Zhan Qi. Expression Recognition Based on Low Pixel Face Images[J]. Laser & Optoelectronics Progress, 2020, 57(10): 101008

    Download Citation

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

    Category: Image Processing

    Received: Aug. 21, 2019

    Accepted: Oct. 18, 2019

    Published Online: May. 8, 2020

    The Author Email: Maojun Li (591338413@qq.com)

    DOI:10.3788/LOP57.101008

    Topics