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

Chinese Journal of Liquid Crystals and Displays, Volume. 40, Issue 6, 881(2025)

Reconstruction method of video snapshot compressive imaging based on Mamba-2

Dunpan SHI1,2,3,4, Wei XU1,3,4、*, Yongjie PIAO1,3,4, Yinghong FANG1,3,4, Haolin JI1,3,4, and Pengfei LI1,2,3,4
Author Affiliations
  • 1Changchun Institute of Optics， Fine Mechanics and Physics， Chinese Academy of Sciences， Changchun 130033， China
  • 2University of Chinese Academy of Sciences， Beijing 100049， China
  • 3Key Laboratory of Space-based Dynamic & Rapid Optical Imaging Technology， Chinese Academy of Sciences， Changchun 130033， China
  • 4Jilin Provincial Key Laboratory of Aerospace Advanced Optical Imaging Technology， Changchun 130033， China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (14)
    Equations (0)
    References (44)
    Tools
    Paper Information
    Topics
    Figures & Tables(14)
    Composition of the video SCI system

    Fig. 1. Composition of the video SCI system

    Download full size
    View in Article
    Structures of Mamba-1 and Mamba-2 module

    Fig. 2. Structures of Mamba-1 and Mamba-2 module

    Download full size
    View in Article
    Architecture diagram of M2BA-SCI

    Fig. 3. Architecture diagram of M2BA-SCI

    Download full size
    View in Article
    Token generation block and video reconstruction block

    Fig. 4. Token generation block and video reconstruction block

    Download full size
    View in Article
    Network of the Mamba-2 linear attention block

    Fig. 5. Network of the Mamba-2 linear attention block

    Download full size
    View in Article
    Grouping tesnet feed forward network［23］

    Fig. 6. Grouping tesnet feed forward network［23］

    Download full size
    View in Article
    Comparison chart of reconstructed video frames on the grayscale simulated video test dataset

    Fig. 7. Comparison chart of reconstructed video frames on the grayscale simulated video test dataset

    Download full size
    View in Article
    Comparison chart of reconstructed video frames on the color simulated video test dataset

    Fig. 8. Comparison chart of reconstructed video frames on the color simulated video test dataset

    Download full size
    View in Article
    Comparison chart of reconstructed video frames of real video data Duomino

    Fig. 9. Comparison chart of reconstructed video frames of real video data Duomino

    Download full size
    View in Article
    Comparison chart of reconstructed video frames of real video data Water Ballon

    Fig. 10. Comparison chart of reconstructed video frames of real video data Water Ballon

    Download full size
    View in Article

    • Table 1. PSNR， SSIM values and running time of different reconstruction algorithms on the grayscale simulated video test dataset

      View table
      View in Article

      Table 1. PSNR， SSIM values and running time of different reconstruction algorithms on the grayscale simulated video test dataset

      DatasetEvaluationGrayscale simulated video test datasetRunning time/s
      AerialCrashDropKobeRunnerTrafficAverage
      PnP-FastDVDnetPSNR SSIM27.870.89526.330.91541.920.98932.330.94336.140.96226.170.91731.790.9376.20
      PnP-FFDNetPSNR SSIM24.290.82024.660.83739.690.98730.320.92332.440.93423.870.82929.210.8880.70
      GAP-TVPSNR SSIM25.020.82624.630.82634.490.96726.640.84030.130.91420.650.69726.930.8450.74
      GAP-CCoTPSNR SSIM29.400.92328.520.94142.540.99232.580.94939.120.98029.030.93833.530.9540.04
      BIRNATPSNR SSIM28.990.91827.840.92742.280.99232.710.95138.700.97729.330.94333.310.9510.11
      RevSCIPSNR SSIM29.350.92528.130.93642.920.992

      33.72

      0.957

      		39.400.97830.020.95033.920.9560.17
      U-netPSNR SSIM27.940.88626.970.89438.130.96129.140.89334.930.957

      24.94

      0.849

      		30.340.9070.02
      EfficientsciPSNR SSIM30.320.93729.270.95443.560.993

      33.45

      0.960

      		39.510.98129.200.94234.220.9610.95
      M2BA-SCI（Ours）PSNR SSIM30.740.94330.280.96243.310.99233.900.961

      40.50

      0.984

      		30.370.95334.850.9660.23

    • Table 2. PSNR， SSIM values and running time of different reconstruction algorithms on the color simulated video test dataset

      View table
      View in Article

      Table 2. PSNR， SSIM values and running time of different reconstruction algorithms on the color simulated video test dataset

      DatasetEvaluationColor simulated video test dataseRunning time/s
      BeautyBosphorusJockeyRunnerShakeNDryTrafficAverage
      PnP-FastDVDnet-grayPSNRSSIM34.580.968

      34.03

      0.953

      33.78

      0.932

      		35.210.93030.400.887

      24.47

      0.832

      		32.080.91726.01
      PnP-FastDVDnet-colorPSNRSSIM

      36.26

      0.976

      		37.100.97536.270.95639.850.97433.890.947

      28.40

      0.924

      		35.310.959102.22
      PnP-FFDNet-grayPSNRSSIM33.150.962

      28.30

      0.897

      		32.310.92030.630.883

      27.72

      0.844

      20.73

      0.695

      		28.810.8673.98
      PnP-FFDNet-colorPSNRSSIM34.600.97033.340.95635.210.94835.490.94132.650.93924.780.84132.680.93254.89
      GAP-TVPSNRSSIM

      33.46

      0.966

      29.60

      0.906

      		29.490.89029.690.874

      29.83

      0.885

      		19.400.61128.580.8553.11
      STFormerPSNRSSIM36.830.98038.360.98837.090.96340.560.98034.670.95229.000.92336.090.9632.96
      M2BA-SCI（ours）PSNRSSIM37.060.97939.340.98536.800.96140.950.98034.440.949

      28.69

      0.922

      		36.210.9631.03

    • Table 3. Ablation experiments of M2BA-SCI on six benchmark grayscale video datasets

      View table
      View in Article

      Table 3. Ablation experiments of M2BA-SCI on six benchmark grayscale video datasets

      实验MLABMamba-2M2FFNEinFFTParams/MFLOPs/GPSNRSSIMRunning time/s
      1√5.16807.3934.320.9620.47
      2√5.17808.1234.400.9630.20
      3√√5.17808.7034.050.9610.38
      4√√5.7882.8934.850.9660.23

    • Table 4. Reconstruction quality and running time on 6 grayscale benchmark datasets using M2BA-SCI with different number of channels and blocks

      View table
      View in Article

      Table 4. Reconstruction quality and running time on 6 grayscale benchmark datasets using M2BA-SCI with different number of channels and blocks

      实验ChannelBlockPSNRSSIMTime/s
      164232.550.9430.12
      2128234.850.9660.23
      3128434.970.9680.64
    Tools

    Get Citation

    Copy Citation Text

    Dunpan SHI, Wei XU, Yongjie PIAO, Yinghong FANG, Haolin JI, Pengfei LI. Reconstruction method of video snapshot compressive imaging based on Mamba-2[J]. Chinese Journal of Liquid Crystals and Displays, 2025, 40(6): 881

    Download Citation

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

    Category:

    Received: Dec. 27, 2024

    Accepted: --

    Published Online: Jul. 14, 2025

    The Author Email: Wei XU (xwciomp@126.com)

    DOI:10.37188/CJLCD.2024-0356

    Topics