LcwtNet: Lifting complex wavelet layers for constructing a compact DNN model

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

    Abstract

    In this paper, a new compact deep neural network (DNN) architecture based on lifting complex wavelets is proposed. The proposed DNN architecture (LcwtNet) is composed of multiple layers in addition to a CNN architecture. Complex wavelet and lifting wavelet layers are introduced as the lower layers of LcwtNet, which can reduce the number of parameters while maintaining high performance similar to that of CNN models. In simulations, the effectiveness of LcwtNet is demonstrated by several test results using the MNIST dataset.

    Original languageEnglish
    Title of host publication2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    Pages288-293
    Number of pages6
    ISBN (Electronic)9781538646625
    DOIs
    Publication statusPublished - Jun 18 2018
    Event17th IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017 - Bilbao, Spain
    Duration: Dec 18 2017Dec 20 2017

    Publication series

    Name2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017

    Other

    Other17th IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017
    CountrySpain
    CityBilbao
    Period12/18/1712/20/17

    Fingerprint

    Network architecture
    Deep neural networks

    All Science Journal Classification (ASJC) codes

    • Safety, Risk, Reliability and Quality
    • Energy Engineering and Power Technology
    • Computer Networks and Communications
    • Computer Vision and Pattern Recognition
    • Hardware and Architecture
    • Signal Processing

    Cite this

    Takano, S. (2018). LcwtNet: Lifting complex wavelet layers for constructing a compact DNN model. In 2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017 (pp. 288-293). (2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ISSPIT.2017.8388657

    LcwtNet : Lifting complex wavelet layers for constructing a compact DNN model. / Takano, Shigeru.

    2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017. Institute of Electrical and Electronics Engineers Inc., 2018. p. 288-293 (2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017).

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

    Takano, S 2018, LcwtNet: Lifting complex wavelet layers for constructing a compact DNN model. in 2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017. 2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017, Institute of Electrical and Electronics Engineers Inc., pp. 288-293, 17th IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017, Bilbao, Spain, 12/18/17. https://doi.org/10.1109/ISSPIT.2017.8388657
    Takano S. LcwtNet: Lifting complex wavelet layers for constructing a compact DNN model. In 2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017. Institute of Electrical and Electronics Engineers Inc. 2018. p. 288-293. (2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017). https://doi.org/10.1109/ISSPIT.2017.8388657
    Takano, Shigeru. / LcwtNet : Lifting complex wavelet layers for constructing a compact DNN model. 2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 288-293 (2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017).
    @inproceedings{20936aa5e425438fb16890aef7507964,
    title = "LcwtNet: Lifting complex wavelet layers for constructing a compact DNN model",
    abstract = "In this paper, a new compact deep neural network (DNN) architecture based on lifting complex wavelets is proposed. The proposed DNN architecture (LcwtNet) is composed of multiple layers in addition to a CNN architecture. Complex wavelet and lifting wavelet layers are introduced as the lower layers of LcwtNet, which can reduce the number of parameters while maintaining high performance similar to that of CNN models. In simulations, the effectiveness of LcwtNet is demonstrated by several test results using the MNIST dataset.",
    author = "Shigeru Takano",
    year = "2018",
    month = "6",
    day = "18",
    doi = "10.1109/ISSPIT.2017.8388657",
    language = "English",
    series = "2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017",
    publisher = "Institute of Electrical and Electronics Engineers Inc.",
    pages = "288--293",
    booktitle = "2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017",
    address = "United States",

    }

    TY - GEN

    T1 - LcwtNet

    T2 - Lifting complex wavelet layers for constructing a compact DNN model

    AU - Takano, Shigeru

    PY - 2018/6/18

    Y1 - 2018/6/18

    N2 - In this paper, a new compact deep neural network (DNN) architecture based on lifting complex wavelets is proposed. The proposed DNN architecture (LcwtNet) is composed of multiple layers in addition to a CNN architecture. Complex wavelet and lifting wavelet layers are introduced as the lower layers of LcwtNet, which can reduce the number of parameters while maintaining high performance similar to that of CNN models. In simulations, the effectiveness of LcwtNet is demonstrated by several test results using the MNIST dataset.

    AB - In this paper, a new compact deep neural network (DNN) architecture based on lifting complex wavelets is proposed. The proposed DNN architecture (LcwtNet) is composed of multiple layers in addition to a CNN architecture. Complex wavelet and lifting wavelet layers are introduced as the lower layers of LcwtNet, which can reduce the number of parameters while maintaining high performance similar to that of CNN models. In simulations, the effectiveness of LcwtNet is demonstrated by several test results using the MNIST dataset.

    UR - http://www.scopus.com/inward/record.url?scp=85050074811&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85050074811&partnerID=8YFLogxK

    U2 - 10.1109/ISSPIT.2017.8388657

    DO - 10.1109/ISSPIT.2017.8388657

    M3 - Conference contribution

    AN - SCOPUS:85050074811

    T3 - 2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017

    SP - 288

    EP - 293

    BT - 2017 IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2017

    PB - Institute of Electrical and Electronics Engineers Inc.

    ER -