Laser & Optoelectronics Progress, Volume. 61, Issue 8, 0811008(2024)

Reconstruction-Free Object Recognition Scheme in Lensless Imaging Systems

Kaiyu Chen1,2,3,4,5, Ying Li1,2,3,4, Zhengdai Li1,2,3,4, and Youming Guo1,2,3,4、*
Author Affiliations
  • 1Key Laboratory on Adaptive Optics, Chinese Academy of Sciences, Chengdu 610209, Sichuan , China
  • 2Institute of Optics and Electronics, Chinese Academy of Sciences, Chengdu 610209, Sichuan , China
  • 3University of Chinese Academy of Sciences, Beijing 100049, China
  • 4School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing 100049, China
  • 5National Key Laboratory of Optical Field Manipulation Science and Technology, Chengdu 610209, Sichuan , China
  • show less
    Figures & Tables(10)
    Principle of lensless imaging
    Comparative experimental group of object recognition schemes
    PSF images corresponding to two types of masks
    Some simulation results
    Schematic of the lensless imaging system
    Experimental device and PSF image of the lensless imaging system. (a) Physical drawing of the device; (b) the camera; (c) PSF image of the random phase mask
    Some real encoded images of lensless imaging
    • Table 1. Experimental results on the MNIST dataset

      View table

      Table 1. Experimental results on the MNIST dataset

      Mask typeNetworkExperimentAccuracy /%
      Phase maskResNet-50Exp1(lensed)99.76
      Exp2(reconstructed)99.68
      Exp3(lensless)99.38
      Swin_TExp1(lensed)99.77
      Exp2(reconstructed)99.67
      Exp3(lensless)99.45
      Amplitude maskResNet-50Exp1(lensed)99.76
      Exp2(reconstructed)99.68
      Exp3(lensless)99.51
      Swin_TExp1(lensed)99.76
      Exp2(reconstructed)99.72
      Exp3(lensless)99.46
    • Table 2. Experimental results on the Fashion MNIST dataset

      View table

      Table 2. Experimental results on the Fashion MNIST dataset

      Mask typeNetworkExperimentAccuracy /%
      Phase maskResNet-50Exp1(lensed)95.77
      Exp2(reconstructed)91.37
      Exp3(lensless)91.28
      Swin_TExp1(lensed)95.34
      Exp2(reconstructed)92.26
      Exp3(lensless)90.40
      Amplitude maskResNet-50Exp1(lensed)95.77
      Exp2(reconstructed)92.65
      Exp3(lensless)92.31
      Swin_TExp1(lensed)95.34
      Exp2(reconstructed)92.98
      Exp3(lensless)91.05
    • Table 3. Experimental results on the real MNIST dataset

      View table

      Table 3. Experimental results on the real MNIST dataset

      NetworkExperimentAccuracy /%
      ResNet-50Exp1(lensed)99.43
      Exp2(reconstructed)98.10
      Exp3(lensless)98.05
      Swin_TExp1(lensed)99.32
      Exp2(reconstructed)98.08
      Exp3(lensless)98.06
    Tools

    Get Citation

    Copy Citation Text

    Kaiyu Chen, Ying Li, Zhengdai Li, Youming Guo. Reconstruction-Free Object Recognition Scheme in Lensless Imaging Systems[J]. Laser & Optoelectronics Progress, 2024, 61(8): 0811008

    Download Citation

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

    Category: Imaging Systems

    Received: Mar. 1, 2023

    Accepted: Apr. 12, 2023

    Published Online: Apr. 16, 2024

    The Author Email: Guo Youming (guoyouming@ioe.ac.cn)

    DOI:10.3788/LOP230755

    Topics