Acta Photonica Sinica, Volume. 52, Issue 12, 1210004(2023)

Infrared and Visible Image Fusion Algorithm Based on Feature Optimization and GAN

Shuai HAO... Jiahao LI, Xu MA*, Tian HE, Siyan SUN and Tong LI |Show fewer author(s)
Author Affiliations
  • College of Electrical and Control Engineering,Xi'an University of Science and Technology,Xi'an 710054,China
  • show less
    Figures & Tables(14)
    Block diagram of the proposed algorithm
    LatLRR decomposition results
    Comparison chart before and after optimization
    MSDC-Fem structure diagram
    Attention fusion process
    Feature reconstruction module
    Discriminator structure
    Subjective experimental results comparison
    Objective experimental result comparison
    Loss function curve
    • Table 1. Adaptive optimization of objective function based on CSA

      View table
      View in Article

      Table 1. Adaptive optimization of objective function based on CSA

      n:number of chameleons in search space
      d:spatial dimension
      t:number of iterations,T:max-number of iterations
      Build adaptive optimization image:Io=yIB+ID
      Design objective function for optimizing parameter y
      F=min(LAG+LSD+λLCON)
      Initialize a population of n chameleons within the search space
      yi=lj+r×(uj-lj)
      Evaluate the fitness of each chameleon using the objective function F
      While(t < T)do
      Step1:Search for prey
      for i=1 to n do
      for j=1 to d do
      if riPp then
      yti,j+p1(pti,j-Gtj)r2+p2(Gtj-yti,j) r1 
      else
      yti,j+μlbjsgn(rand-0.5)+ μ[(uj-lj)r3]sgn(rand-0.5)
      end if
      end for
      end for
      Step2:Eyes’ rotation reveals prey
      for i=1 to n do
      yt+1i=m×(yti-y¯ti)+y¯ti
      end for
      Step3:Hunting process for prey
      for i=1 to n do
      for j=1 to d do
      yt+1i,j=yti,j+[(vti,j)2-(vt-1i,j)2]/(2a)
      end for
      end for
      Evaluate the new positions of the chameleons
      Update the position of the chameleons
      Evaluate the fitness of each chameleon
      Return best solution y
      t = t + 1
      end while
      Obtain the global optimal solution y,and then obtain the optimized image Io
    • Table 2. Training process of network model

      View table
      View in Article

      Table 2. Training process of network model

      step1:for M epochs do
      step2:for p times do
      step3:select b visible image samples:{Ivis1,Ivis2,,Ivisb}
      step4:select b infrared image samples:{Iir1,Iir2,,Iirb}
      step5:select b fusion image samples:{Ifused1,Ifused2,,Ifusedb}

      step6:Using the Adam optimizer to update discriminator parameters:

      D(LD=LD-ir+LD-vis)

      step7:end for
      step8:select b visible image samples:{Ivis1,Ivis2,,Ivisb}
      step9:select b infrared image samples:{Iir1,Iir2,,Iirb}

      step10:Using the Adam optimizer to update generator parameters:

      G(LG=Ladv+λ1Lcontent)

      step11:end for
    • Table 3. Average running time of different algorithms(units:s)

      View table
      View in Article

      Table 3. Average running time of different algorithms(units:s)

      Algorithms

      DenseFuse

      FusionGAN

      ResNet-ZCA

      MDLatLRR

      PMGI

      RFN-Nest

      Ours

      Time

      0.124

      1.846

      1.719

      5.846

      0.637

      0.284

      0.451 9

    • Table 4. Ablation experiments objectively results comparison

      View table
      View in Article

      Table 4. Ablation experiments objectively results comparison

      Module AModule BModule CENSFJEVIFSSIMQAB/F
      6.389 55.864 211.956 10.664 20.813 40.307 0
      6.415 86.468 511.990 70.693 40.824 60.350 9
      6.625 87.130 112.128 40.691 50.826 90.377 1
      6.489 66.256 112.034 20.689 80.819 80.315 9
      6.842 97.956 012.362 90.715 30.863 70.388 0
      6.559 36.594 712.186 30.708 80.820 40.375 9
      6.952 48.556 312.461 20.833 50.822 50.411 8
      7.168 810.006 512.650 90.868 60.854 20.457 4
    Tools

    Get Citation

    Copy Citation Text

    Shuai HAO, Jiahao LI, Xu MA, Tian HE, Siyan SUN, Tong LI. Infrared and Visible Image Fusion Algorithm Based on Feature Optimization and GAN[J]. Acta Photonica Sinica, 2023, 52(12): 1210004

    Download Citation

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

    Category:

    Received: May. 10, 2023

    Accepted: Jul. 24, 2023

    Published Online: Feb. 19, 2024

    The Author Email: MA Xu (maxu@xust.edu.cn)

    DOI:10.3788/gzxb20235212.1210004

    Topics