• Acta Optica Sinica
  • Vol. 39, Issue 2, 0210003 (2019)
Zhihong Xi*, Caiyan Hou, Kunpeng Yuan, and Zhuoqun Xue
Author Affiliations
  • College of Information and Communication Engineering, Harbin Engineering University, Harbin, Heilongjiang 150001, China
  • show less
    DOI: 10.3788/AOS201939.0210003 Cite this Article Set citation alerts
    Zhihong Xi, Caiyan Hou, Kunpeng Yuan, Zhuoqun Xue. Super-Resolution Reconstruction of Accelerated Image Based on Deep Residual Network[J]. Acta Optica Sinica, 2019, 39(2): 0210003 Copy Citation Text show less
    Diagram of SRCNN structure
    Fig. 1. Diagram of SRCNN structure
    Diagram of ESPCN structure
    Fig. 2. Diagram of ESPCN structure
    Diagram of network structure of the proposed algorithm
    Fig. 3. Diagram of network structure of the proposed algorithm
    Residual network structure
    Fig. 4. Residual network structure
    (a) Variation of loss function of 12-layer network with number of iterations; (b) variation of PSNR average value of set 5 with number of iterations under different layers
    Fig. 5. (a) Variation of loss function of 12-layer network with number of iterations; (b) variation of PSNR average value of set 5 with number of iterations under different layers
    Variation of PSNR average value of set 5 under different activation functions with number of iterations
    Fig. 6. Variation of PSNR average value of set 5 under different activation functions with number of iterations
    Relationship between running time and PSNR average value of set 5 under different algorithms
    Fig. 7. Relationship between running time and PSNR average value of set 5 under different algorithms
    Variation of PSNR average value of set 5 under different optimization methods with number of iterations
    Fig. 8. Variation of PSNR average value of set 5 under different optimization methods with number of iterations
    Variation of PSNR average value of set 5 under different filter numbers with number of iterations
    Fig. 9. Variation of PSNR average value of set 5 under different filter numbers with number of iterations
    Variation of PSNR average value of set 5 under different network models with number of iterations. (a) Networks of 6-layer and 8-layer; (b) networks of 10-layer and 12-layer
    Fig. 10. Variation of PSNR average value of set 5 under different network models with number of iterations. (a) Networks of 6-layer and 8-layer; (b) networks of 10-layer and 12-layer
    Effect of Monarch under different algorithms
    Fig. 11. Effect of Monarch under different algorithms
    Effect of Comic under different algorithms
    Fig. 12. Effect of Comic under different algorithms
    DepthPSNR average valueSSIM average value
    set 5set 14set 5set 14
    633.4829.650.92490.8936
    1033.5829.670.92650.8941
    1233.6029.690.92680.8944
    Table 1. PSNR/SSIM average values of test sets at different depths
    Data setScaleBicubicScSRNE+LLEANRSRCNNESPCNDRSR
    ×233.6535.1335.7635.8336.3636.3937.41
    set 5×330.4231.5431.9132.0032.5232.7833.60
    ×428.4428.2429.6629.7430.1530.2131.18
    ×230.2131.3631.7831.8132.2132.2132.95
    set 14×327.5128.3628.5928.6429.0329.1329.69
    ×425.9725.9726.7826.8327.2327.1727.83
    ×229.4129.4130.3930.4330.8930.9331.55
    BSD100×327.0727.6727.7827.8128.1128.2728.54
    ×425.8425.8426.3926.4126.6326.5927.00
    Table 2. PSNR average value of set 5, set 14, and BSD100 under different algorithms
    Data setScaleBicubicScSRNE+LLEANRSRCNNESPCNDRSR
    ×20.93550.94280.95370.95460.95660.95680.9621
    set 5×30.87790.88510.90530.90620.91290.91620.9268
    ×40.81850.80250.85160.85330.86210.85780.8846
    ×20.93480.95640.95750.95780.95910.95980.9630
    set 14×30.84940.86490.88020.87770.88460.88740.8941
    ×40.77990.78000.81520.81730.82060.82250.8361
    ×20.84010.87020.87290.87420.88270.88320.8924
    BSD100×30.73010.75140.77020.77180.77780.78160.7917
    ×40.64700.64700.68750.68960.69190.69420.7092
    Table 3. SSIM average value of set 5, set 14, and BSD100 under different algorithms
    Zhihong Xi, Caiyan Hou, Kunpeng Yuan, Zhuoqun Xue. Super-Resolution Reconstruction of Accelerated Image Based on Deep Residual Network[J]. Acta Optica Sinica, 2019, 39(2): 0210003
    Download Citation