Laser & Optoelectronics Progress, Volume. 57, Issue 14, 141007(2020)

Improved Non-Local Mean Denoising Algorithm Based on Difference Hash Algorithm

Chunjian Hua1,2、*, Jinke Ma1,2, and Ying Chen3
Author Affiliations
  • 1School of Mechanical Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • 2Jiangsu Key Laboratory of Advanced Food Manufacturing Equipment & Technology, Wuxi, Jiangsu 214122, China;
  • 3School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • show less
    Figures & Tables(9)
    Flow chart of algorithm
    Test images. (a) Lena; (b) Cameraman; (c) Lena after adding noise; (d) Cameraman after adding noise
    Fitting result of σ' and h of Lena image
    Fitting result of σ' and h of Cameraman image
    Denoising results of different algorithms. (a) NLM; (b) GNLM; (c) TNLM; (d) HNLM
    Image details obtained by different algorithms. (a) NLM; (b) GNLM; (c) TNLM; (d) HNLM
    • Table 1. λ and h when PSNR is maximum

      View table

      Table 1. λ and h when PSNR is maximum

      σImagePSNR /dBλh
      10Lena34.720.50.006
      Cameraman35.510.50.006
      15Lena33.060.50.014
      Cameraman33.500.50.014
      20Lena31.700.50.026
      Cameraman32.040.50.026
      25Lena30.580.50.044
      Cameraman30.710.50.044
      30Lena29.660.50.066
      Cameraman29.580.50.064
      35Lena28.860.50.092
      Cameraman28.430.50.090
    • Table 2. PSNR and running time of four algorithms under different noises (Lena image)

      View table

      Table 2. PSNR and running time of four algorithms under different noises (Lena image)

      AlgorithmParameterStandard deviation
      102550
      NLMPSNR /dB27.3826.9825.71
      time /s0.450.450.46
      GNLMPSNR /dB34.1130.5327.12
      time /s3.223.173.55
      TNLMPSNR /dB33.2229.4025.99
      time/s3.543.453.51
      HNLMPSNR /dB34.7230.5827.02
      time /s0.630.650.62
    • Table 3. PSNR and running time of four algorithms under different noises (Cameraman image)

      View table

      Table 3. PSNR and running time of four algorithms under different noises (Cameraman image)

      AlgorithmParameterStandard deviation
      102550
      NLMPSNR /dB26.0125.5624.02
      time /s0.410.430.42
      GNLMPSNR /dB34.2030.5225.95
      Time /s3.213.253.26
      TNLMPSNR /dB34.1629.8324.25
      time /s3.523.493.34
      HNLMPSNR /dB35.5130.7125.89
      time /s0.650.690.67
    Tools

    Get Citation

    Copy Citation Text

    Chunjian Hua, Jinke Ma, Ying Chen. Improved Non-Local Mean Denoising Algorithm Based on Difference Hash Algorithm[J]. Laser & Optoelectronics Progress, 2020, 57(14): 141007

    Download Citation

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

    Category: Image Processing

    Received: Sep. 9, 2019

    Accepted: Dec. 11, 2019

    Published Online: Jul. 23, 2020

    The Author Email: Chunjian Hua (cjhua@jiangnan.edu.cn)

    DOI:10.3788/LOP57.141007

    Topics