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 Lasers, Volume. 46, Issue 10, 1009001(2019)

End-to-End Multispectral Image Compression Using Convolutional Neural Network

Fanqiang Kong1, Yongbo Zhou1、*, Qiu Shen2, and Keyao Wen1
Author Affiliations
  • 1College of Astronautics, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu 210000, China
  • 2School of Electronic Science and Engineering, Nanjing University, Nanjing, Jiangsu 210046, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (9)
    Equations (0)
    References (20)
    Cited By (10)
    Tools
    Paper Information
    Topics
    Figures & Tables(9)
    Overall flow chart of compression network

    Fig. 1. Overall flow chart of compression network

    Download full size
    Intermediate feature image visualization. (a) Original image; (b) fMap_5; (c) fMap_16; (d) fMap_32; (e) fMap_46

    Fig. 2. Intermediate feature image visualization. (a) Original image; (b) fMap_5; (c) fMap_16; (d) fMap_32; (e) fMap_46

    Download full size
    Illustrations of residual network unit. (a) Default architecture in ResNet; (b) proposed architecture ResUnit by replacing ReLU

    Fig. 3. Illustrations of residual network unit. (a) Default architecture in ResNet; (b) proposed architecture ResUnit by replacing ReLU

    Download full size
    Illustrations of compression networks. (a) Forward coding network architecture; (b) reverse decoding network architecture

    Fig. 4. Illustrations of compression networks. (a) Forward coding network architecture; (b) reverse decoding network architecture

    Download full size
    Average PSNR of test dataset at different code rates

    Fig. 5. Average PSNR of test dataset at different code rates

    Download full size
    RD curves. (a) tj-chun118; (b) hunan-qiu233; (c) ss443; (d) tj-xia216

    Fig. 6. RD curves. (a) tj-chun118; (b) hunan-qiu233; (c) ss443; (d) tj-xia216

    Download full size
    Reconstructed image display. (a) Original image; (b) JPEG; (c) JPEG2000; (d) proposed algorithm

    Fig. 7. Reconstructed image display. (a) Original image; (b) JPEG; (c) JPEG2000; (d) proposed algorithm

    Download full size
    Spectral similarity curves

    Fig. 8. Spectral similarity curves

    Download full size

    • Table 1. Average spectral similarity of test dataset at different code rates

      View table

      Table 1. Average spectral similarity of test dataset at different code rates

      Test datasetCode rate /(bit·pixel-1)
      0.0980.1160.1370.1510.2060.2940.3450.455
      JPEG1429.5601230.880616.868445.511283.592269.364187.519164.105
      JPEG2000396.171312.923282.149271.791260.262248.623154.931149.086
      Proposed189.911185.021178.488174.609157.862149.392109.81193.105
    Tools

    Get Citation

    Copy Citation Text

    Fanqiang Kong, Yongbo Zhou, Qiu Shen, Keyao Wen. End-to-End Multispectral Image Compression Using Convolutional Neural Network[J]. Chinese Journal of Lasers, 2019, 46(10): 1009001

    Download Citation

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

    Category: holography and information processing

    Received: May. 22, 2019

    Accepted: Jun. 17, 2019

    Published Online: Oct. 25, 2019

    The Author Email: Zhou Yongbo (zybfight@163.com)

    DOI:10.3788/CJL201946.1009001

    Topics

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