Laser & Optoelectronics Progress, Volume. 58, Issue 6, 610004(2021)

X-Ray Image Reconstruction Based on Hierarchical Model and Low-Rank Approximation

Wang Jiayu, Xu Jinxin, and Li Qingwu*
Author Affiliations
  • College of Internet of Things Engineering, Hohai University, Changzhou, Jiangsu 213022, China
  • show less
    Figures & Tables(6)
    Reconstruction results of FTO variable density image at 3% noise level. (a) Noise-containing optical path image; (b) original image; (c) GPSR; (d) PNM_WLSA; (e) LRIS_ Gamma; (f) LRIS_ Jefferys; (g) SPA; (h) proposed method
    Reconstruction results of FTO constant density image at 3% noise level. (a) Noise-containing optical path image; (b) original image; (c) GPSR; (d) PNM_WLSA; (e) LRIS_ Gamma; (f) LRIS_ Jefferys; (g) SPA; (h) proposed method
    Real X-ray images and reconstructed results. (a) Real X-ray images; (b) reconstruction results
    Reconstruction results of different algorithms. (a) Regional profile diagram of tin column (CL3); (b) regional profile diagram of tin disk (DR1)
    • Table 1. RE index comparison of reconstruction results

      View table

      Table 1. RE index comparison of reconstruction results

      LevelImageGPSRPNM_WLSALRIS_GammaLRIS_JefferySPAProposed method
      WithoutVariable4.55551.03472.59440.59690.19050.0314
      Constant4.79990.06471.48130.53060.05750.0167
      1%Variable12.587210.50307.50387.415912.43556.6443
      Constant11.515010.32937.60237.583012.12136.7618
      3%Variable30.087618.459011.039410.965837.305410.5727
      Constant34.771221.225611.263011.419336.367410.9413
      Average16.386110.26946.91406.418616.41295.8280
    • Table 2. Inversion density mean and RE values of real X-ray images

      View table

      Table 2. Inversion density mean and RE values of real X-ray images

      MaterialTypeGPSRLRIS_GammaLRIS_JefferySPAProposed method
      Mean value /(g·cm-3)RE /%Mean value /(g·cm-3)RE /%Mean value /(g·cm-3)RE /%Mean value /(g·cm-3)RE /%Mean value /(g·cm-3)RE /%
      Tin columnCR18.013911.74017.747511.74677.748311.73667.715510.80967.62907.3034
      CR27.83439.11347.881210.78347.881110.77617.808910.74687.73407.2611
      CR37.52266.09887.59959.69787.59979.70527.56026.47327.37165.1178
      CL17.53486.99727.58239.87357.58149.87837.58888.31827.53336.0993
      CL27.67637.34147.74439.93317.74419.93477.70038.23237.65516.6236
      CL37.38305.88347.39967.34847.39937.33757.36006.29297.28704.1236
      Tin diskDR12.760929.62242.787836.28602.788236.29802.758030.58132.680020.5984
      DR22.648235.71422.653348.13082.653748.16132.627739.34472.603127.5855
      DL12.689321.53672.712931.27932.712831.29822.693422.36982.608217.7010
      DL22.640328.43982.698337.84502.698137.84842.693331.50252.600221.6413
    Tools

    Get Citation

    Copy Citation Text

    Wang Jiayu, Xu Jinxin, Li Qingwu. X-Ray Image Reconstruction Based on Hierarchical Model and Low-Rank Approximation[J]. Laser & Optoelectronics Progress, 2021, 58(6): 610004

    Download Citation

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

    Category: Image Processing

    Received: Jul. 31, 2020

    Accepted: --

    Published Online: Mar. 11, 2021

    The Author Email: Qingwu Li (li_qingwu@163.com)

    DOI:10.3788/LOP202158.0610004

    Topics