Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 59, Issue 22, 2215007(2022)

Image Deblurring Based on Enhanced Multiscale Feature Network

Zhijun Yu, Guodong Wang*, and Xinyue Zhang
Author Affiliations
  • College of Computer Science & Technology, Qingdao University, Qingdao 266071, Shandong, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (8)
    Equations (6)
    References (40)
    Cited By (4)
    Tools
    Paper Information
    Topics
    Figures & Tables(8)
    Illustration of proposed image deblurring method based on enhanced multi-scale feature network

    Fig. 1. Illustration of proposed image deblurring method based on enhanced multi-scale feature network

    Download full size
    Enhanced multi-scale residual feature extracted module

    Fig. 2. Enhanced multi-scale residual feature extracted module

    Download full size
    Illustration of the reconstruction module and the cross-stage feature fusion module. (a) Reconstruction module; (b) cross-stage feature fusion module

    Fig. 3. Illustration of the reconstruction module and the cross-stage feature fusion module. (a) Reconstruction module; (b) cross-stage feature fusion module

    Download full size
    Cross-stage attention architecture

    Fig. 4. Cross-stage attention architecture

    Download full size
    Visualization of deblurring results on the GoPro dataset

    Fig. 5. Visualization of deblurring results on the GoPro dataset

    Download full size
    Visualization of deblurring results on the HIDE dataset

    Fig. 6. Visualization of deblurring results on the HIDE dataset

    Download full size

    • Table 1. Deblurring results of different methods on GoPro dataset and HIDE dataset

      View table

      Table 1. Deblurring results of different methods on GoPro dataset and HIDE dataset

      MethodConferenceGoProHIDE
      PSNR /dBSSIMPSNR /dBSSIM
      Method in Ref.［10］CVPR201423.540.822
      Method in Ref.［26］CVPR201931.200.94029.090.924
      Method in Ref.［27］CVPR202132.660.95930.960.939
      Method in Ref.［29］ICCV201929.550.93426.160.875
      Method in Ref.［33］CVPR201830.260.93428.360.915
      Method in Ref.［38］CVPR202029.980.93029.980.930
      Proposed method32.870.96731.030.942

    • Table 2. Results of ablation experiment

      View table

      Table 2. Results of ablation experiment

      MethodNumber of MSRBCSACSFFLossPSNR /dBSSIM
      Base model4√√√32.870.967
      EMF-block44×√√31.350.944
      EMF-block44√×√30.500.931
      EMF-block44√√×32.530.960
      EMF-block66√√√32.880.967
      EMF-block88√√√32.500.959
    Tools

    Get Citation

    Copy Citation Text

    Zhijun Yu, Guodong Wang, Xinyue Zhang. Image Deblurring Based on Enhanced Multiscale Feature Network[J]. Laser & Optoelectronics Progress, 2022, 59(22): 2215007

    Download Citation

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

    Category: Machine Vision

    Received: Aug. 27, 2021

    Accepted: Oct. 13, 2021

    Published Online: Oct. 13, 2022

    The Author Email: Guodong Wang (doctorwgd@gmail.com)

    DOI:10.3788/LOP202259.2215007

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology