• Laser & Optoelectronics Progress
  • Vol. 55, Issue 11, 111010 (2018)
Min Wang, Jing Hao*, Chenhong Yao, and Qiqi Shi
Author Affiliations
  • School of Information and Control Engineering, Xi'an University of Architecture and Technology, Xi'an, Shaanxi 710055, China
  • show less
    DOI: 10.3788/LOP55.111010 Cite this Article Set citation alerts
    Min Wang, Jing Hao, Chenhong Yao, Qiqi Shi. Sign Language Semantic Recognition Based on Optimized Fully Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2018, 55(11): 111010 Copy Citation Text show less
    Self-acquired gesture images
    Fig. 1. Self-acquired gesture images
    Network diagram of gesture recognition algorithm
    Fig. 2. Network diagram of gesture recognition algorithm
    Error rate curve of gesture recognition
    Fig. 3. Error rate curve of gesture recognition
    Noised images. (a) Gaussian noise with variance of 0.01; (b) Gaussian noise with variance of 0.02; (c) Gaussian noise with variance of 0.03
    Fig. 4. Noised images. (a) Gaussian noise with variance of 0.01; (b) Gaussian noise with variance of 0.02; (c) Gaussian noise with variance of 0.03
    ParameterlayerKernal size /(pixel×pixel)Input /(pixel×pixel)
    C17×758×58@6
    dS12×229×29@6
    C26×624×24@8
    dS22×212×12@8
    C35×58×8@10
    dS32×24×4@10
    C43×32×2@12
    uS12×24×4@12
    uS22×28×8@12
    uS38×864×64@12
    Table 1. Parameters of fully convolutional neural network
    ItemGesture category
    STUVW
    Recognition rate96.9296.8597.8697.9397.50
    Table 2. Gesture recognition rate by FCN algorithm%
    Gaussian noisevarianceGesture category
    STUVW
    0.0196.7496.7297.3497.3797.18
    0.0296.3196.2797.0497.3196.87
    0.0396.1596.0296.8397.0396.53
    Table 3. Recognition rate by FCN algorithm under different Gaussian noises%
    AlgorithmGesture category
    STUVW
    Algorithm in Ref. [4]95.8695.7695.8396.0396.12
    Algorithm in Ref. [6]95.9496.1896.2296.3896.35
    Algorithm in Ref. [8]96.2196.4396.3296.8697.15
    FCN algorithm96.9296.8597.8697.9397.50
    Table 4. Performance comparison between FCN and other algorithms%
    Min Wang, Jing Hao, Chenhong Yao, Qiqi Shi. Sign Language Semantic Recognition Based on Optimized Fully Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2018, 55(11): 111010
    Download Citation