Infrared and Laser Engineering, Volume. 50, Issue 11, 20210021(2021)

Light field depth estimation of fusing consistency and difference constraints

Zeyang He1...2, Huiping Deng1,2, Sen Xiang1,2, and Jin Wu12 |Show fewer author(s)
Author Affiliations
  • 1School of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan 430081, China
  • 2Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education, Wuhan University of Science and Technology, Wuhan 430081, China
  • show less
    Figures & Tables(11)
    Schematic diagram of light field and EPI acquisition (top), and diagram after overlaying light field images (bottom)
    Schematic diagram of SPO and SLO
    The flow chart of light field depth estimation
    Analysis of the accuracy of the cost function
    Comparison of the accuracy of the two cost functions
    Comparison of experimental results of depth maps with two cost functions
    Comparison of the results of the depth bad pixel map
    Comparison of boxes and dino depth map detail results
    • Table 1. Physical significance of important symbols

      View table
      View in Article

      Table 1. Physical significance of important symbols

      SymbolPhysical significanceSymbolPhysical significance
      R(x*,θN) EPI slashwθ(i,j) SPO distance weight
      E(x*+ $\Delta $x N,u) EPI slash pixel coordinate informationχ2(aθ,bθ) SPO histogram χ2 distance
      tanθNEPI slopeDs(P,θN) SPO cost function
      I1,I2Pixel adjacent integer pixel valuec,cs,ceAdaptive confidence weight
      w1,w2Horizontal distance weightD(P,θN) One-way final cost function
      h(x,u′) Pixel value probabilityDu,v(x,y,θ) Final cost function
      Cei(P,θN) Initial single channel entropy costEunaryMRF data cost
      w(P,θN) Color similarity weightES(p,p(α)) SPO data cost
      Ce(P,θN) Final single channel entropy costEe(p,p(α)) SLO data cost
      De(P,θN) SLO color entropy cost functionEbinaryMRF smoothing term
    • Table 2. Comparison of results of BP>0.07

      View table
      View in Article

      Table 2. Comparison of results of BP>0.07

      BP>0.07
      boxesCottonSideboarddino
      CAE0.187 80.038 40.120 00.068 8
      IGF0.192 40.141 80.108 30.070 5
      LF0.287 40.100 10.243 20.210 9
      LF_DC0.893 30.889 70.957 80.869 1
      LF_OOC0.405 60.365 50.436 40.252 7
      LF_PAC0.225 50.073 70.100 50.091 0
      MBM0.201 30.068 40.163 30.082 9
      POBR0.397 80.120 60.296 20.271 8
      SPO0.143 30.031 80.073 30.025 1
      Proposed0.141 10.025 50.071 90.023 7
    • Table 3. Comparison of results of MSE

      View table
      View in Article

      Table 3. Comparison of results of MSE

      MSE
      boxesCottonSideboarddino
      CAE0.089 10.026 90.008 60.004 0
      IGF0.106 10.108 30.012 00.009 9
      LF0.162 90.141 00.039 00.017 1
      LF_DC0.121 20.061 80.575 50.030 5
      LF_OOC0.074 80.056 70.043 00.017 2
      LF_PAC0.099 20.070 50.011 00.009 2
      MBM0.101 90.067 10.029 10.013 7
      POBR0.134 80.063 50.046 80.023 4
      SPO0.108 80.041 40.009 90.003 9
      Proposed0.101 70.024 80.009 60.003 5
    Tools

    Get Citation

    Copy Citation Text

    Zeyang He, Huiping Deng, Sen Xiang, Jin Wu. Light field depth estimation of fusing consistency and difference constraints[J]. Infrared and Laser Engineering, 2021, 50(11): 20210021

    Download Citation

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

    Category: Image processing

    Received: May. 25, 2021

    Accepted: --

    Published Online: Dec. 7, 2021

    The Author Email:

    DOI:10.3788/IRLA20210021

    Topics