Deep Inverse Tone Mapping Optimized for High Dynamic Range Display

Katsuhiko Hirao, Zhengxue Cheng, Masaru Takeuchi, Jiro Katto

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    The popularity of high dynamic range (HDR) makes the inverse tone mapping become an important technique for HDR display. In this paper, we propose a convolutional neural network (CNN) based inverse tone mapping method to generate a high-quality HDR image from one single standard dynamic range (SDR) image. First, we present a CNN design with a three-channel input, which considers both luminance and chrominance. Second, we propose to use overlapped inputs to remove the boundary artifacts, caused by zero padding in CNN. Experimental results demonstrate the high quality of our generated HDR images compared to the ground truth and conventional inverse tone mapping methods.

    Original languageEnglish
    Title of host publication9th International Conference on Information and Communication Technology Convergence
    Subtitle of host publicationICT Convergence Powered by Smart Intelligence, ICTC 2018
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    Pages472-474
    Number of pages3
    ISBN (Electronic)9781538650400
    DOIs
    Publication statusPublished - 2018 Nov 16
    Event9th International Conference on Information and Communication Technology Convergence, ICTC 2018 - Jeju Island, Korea, Republic of
    Duration: 2018 Oct 172018 Oct 19

    Other

    Other9th International Conference on Information and Communication Technology Convergence, ICTC 2018
    CountryKorea, Republic of
    CityJeju Island
    Period18/10/1718/10/19

    Fingerprint

    Display devices
    Neural networks
    Luminance

    Keywords

    • convolutional neural networks
    • high dynamic range imaging
    • inverse tone mapping

    ASJC Scopus subject areas

    • Computer Networks and Communications
    • Computer Science Applications
    • Information Systems
    • Information Systems and Management
    • Artificial Intelligence

    Cite this

    Hirao, K., Cheng, Z., Takeuchi, M., & Katto, J. (2018). Deep Inverse Tone Mapping Optimized for High Dynamic Range Display. In 9th International Conference on Information and Communication Technology Convergence: ICT Convergence Powered by Smart Intelligence, ICTC 2018 (pp. 472-474). [8539648] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICTC.2018.8539648

    Deep Inverse Tone Mapping Optimized for High Dynamic Range Display. / Hirao, Katsuhiko; Cheng, Zhengxue; Takeuchi, Masaru; Katto, Jiro.

    9th International Conference on Information and Communication Technology Convergence: ICT Convergence Powered by Smart Intelligence, ICTC 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 472-474 8539648.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Hirao, K, Cheng, Z, Takeuchi, M & Katto, J 2018, Deep Inverse Tone Mapping Optimized for High Dynamic Range Display. in 9th International Conference on Information and Communication Technology Convergence: ICT Convergence Powered by Smart Intelligence, ICTC 2018., 8539648, Institute of Electrical and Electronics Engineers Inc., pp. 472-474, 9th International Conference on Information and Communication Technology Convergence, ICTC 2018, Jeju Island, Korea, Republic of, 18/10/17. https://doi.org/10.1109/ICTC.2018.8539648
    Hirao K, Cheng Z, Takeuchi M, Katto J. Deep Inverse Tone Mapping Optimized for High Dynamic Range Display. In 9th International Conference on Information and Communication Technology Convergence: ICT Convergence Powered by Smart Intelligence, ICTC 2018. Institute of Electrical and Electronics Engineers Inc. 2018. p. 472-474. 8539648 https://doi.org/10.1109/ICTC.2018.8539648
    Hirao, Katsuhiko ; Cheng, Zhengxue ; Takeuchi, Masaru ; Katto, Jiro. / Deep Inverse Tone Mapping Optimized for High Dynamic Range Display. 9th International Conference on Information and Communication Technology Convergence: ICT Convergence Powered by Smart Intelligence, ICTC 2018. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 472-474
    @inproceedings{e02374cce15246b88e7f40eef2ad85de,
    title = "Deep Inverse Tone Mapping Optimized for High Dynamic Range Display",
    abstract = "The popularity of high dynamic range (HDR) makes the inverse tone mapping become an important technique for HDR display. In this paper, we propose a convolutional neural network (CNN) based inverse tone mapping method to generate a high-quality HDR image from one single standard dynamic range (SDR) image. First, we present a CNN design with a three-channel input, which considers both luminance and chrominance. Second, we propose to use overlapped inputs to remove the boundary artifacts, caused by zero padding in CNN. Experimental results demonstrate the high quality of our generated HDR images compared to the ground truth and conventional inverse tone mapping methods.",
    keywords = "convolutional neural networks, high dynamic range imaging, inverse tone mapping",
    author = "Katsuhiko Hirao and Zhengxue Cheng and Masaru Takeuchi and Jiro Katto",
    year = "2018",
    month = "11",
    day = "16",
    doi = "10.1109/ICTC.2018.8539648",
    language = "English",
    pages = "472--474",
    booktitle = "9th International Conference on Information and Communication Technology Convergence",
    publisher = "Institute of Electrical and Electronics Engineers Inc.",

    }

    TY - GEN

    T1 - Deep Inverse Tone Mapping Optimized for High Dynamic Range Display

    AU - Hirao, Katsuhiko

    AU - Cheng, Zhengxue

    AU - Takeuchi, Masaru

    AU - Katto, Jiro

    PY - 2018/11/16

    Y1 - 2018/11/16

    N2 - The popularity of high dynamic range (HDR) makes the inverse tone mapping become an important technique for HDR display. In this paper, we propose a convolutional neural network (CNN) based inverse tone mapping method to generate a high-quality HDR image from one single standard dynamic range (SDR) image. First, we present a CNN design with a three-channel input, which considers both luminance and chrominance. Second, we propose to use overlapped inputs to remove the boundary artifacts, caused by zero padding in CNN. Experimental results demonstrate the high quality of our generated HDR images compared to the ground truth and conventional inverse tone mapping methods.

    AB - The popularity of high dynamic range (HDR) makes the inverse tone mapping become an important technique for HDR display. In this paper, we propose a convolutional neural network (CNN) based inverse tone mapping method to generate a high-quality HDR image from one single standard dynamic range (SDR) image. First, we present a CNN design with a three-channel input, which considers both luminance and chrominance. Second, we propose to use overlapped inputs to remove the boundary artifacts, caused by zero padding in CNN. Experimental results demonstrate the high quality of our generated HDR images compared to the ground truth and conventional inverse tone mapping methods.

    KW - convolutional neural networks

    KW - high dynamic range imaging

    KW - inverse tone mapping

    UR - http://www.scopus.com/inward/record.url?scp=85059462692&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85059462692&partnerID=8YFLogxK

    U2 - 10.1109/ICTC.2018.8539648

    DO - 10.1109/ICTC.2018.8539648

    M3 - Conference contribution

    SP - 472

    EP - 474

    BT - 9th International Conference on Information and Communication Technology Convergence

    PB - Institute of Electrical and Electronics Engineers Inc.

    ER -