Laser & Optoelectronics Progress, Volume. 58, Issue 4, 0410013(2021)

Image Super-Resolution Reconstruction Method Based on Self-Attention Deep Network

Zihan Chen1, Haobo Wu2、*, Haodong Pei3、*, Rong Chen1, Jiaxin Hu1, and Hengtong Shi1
Author Affiliations
  • 1Futian Power Supply Bureau, Shenzhen Power Supply Bureau Co., Ltd., Shenzhen 518001, China
  • 2School of Electronic Engineering, Xidian University, Xi'an, Shaanxi 710071, China
  • 3Key Laboratory of Intelligent Infrared Perception, Shanghai Institute of Technical Physics, Chinese Academy of Sciences, Shanghai 200083, China
  • show less
    Figures & Tables(7)
    Overall structure of SADeepNet
    Structure of self-attention layer
    Plane images reconstructed by different methods. (a) Low-resolution input image with enlarged display; (b) Bicubic; (c) SRCNN; (d) FSRCNN; (e) SADeepNet; (f) high-resolution original image
    Chair images reconstructed by different methods. (a) Low-resolution input image with enlarged display; (b) Bicubic; (c) SRCNN; (d) FSRCNN; (e) SADeepNet; (f) high-resolution original image
    Butterfly images reconstructed by different methods. (a) Low-resolution input image with enlarged display; (b) Bicubic; (c) SRCNN; (d) FSRCNN; (e) SADeepNet; (f) high-resolution original image
    • Table 1. PSNR comparison of different methods unit: dB

      View table

      Table 1. PSNR comparison of different methods unit: dB

      DatasetBicubicSRCNNFSRCNNSADeepNet
      Plane(×2)19.9921.1621.3722.10
      Plane(×3)18.3919.2120.0821.22
      Chair(×2)18.4018.5518.9619.67
      Chair(×3)17.3217.9218.0418.46
      Butterfly(×2)16.3816.8117.1317.91
      Butterfly(×3)16.2316.7217.0717.75
    • Table 2. SSIM comparison of different methods

      View table

      Table 2. SSIM comparison of different methods

      DatasetBicubicSRCNNFSRCNNSADeepNet
      Plane(×2)0.660.700.710.73
      Plane(×3)0.620.640.660.69
      Chair(×2)0.550.570.590.63
      Chair(×3)0.520.520.550.58
      Butterfly(×2)0.560.570.590.64
      Butterfly(×3)0.530.550.580.60
    Tools

    Get Citation

    Copy Citation Text

    Zihan Chen, Haobo Wu, Haodong Pei, Rong Chen, Jiaxin Hu, Hengtong Shi. Image Super-Resolution Reconstruction Method Based on Self-Attention Deep Network[J]. Laser & Optoelectronics Progress, 2021, 58(4): 0410013

    Download Citation

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

    Category: Image Processing

    Received: Jun. 30, 2020

    Accepted: Aug. 7, 2020

    Published Online: Feb. 25, 2021

    The Author Email: Wu Haobo (haobow@126.com), Pei Haodong (haobow@126.com)

    DOI:10.3788/LOP202158.0410013

    Topics