中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

62 records (took 0.026 seconds)

    METHOD AND APPARATUS FOR RENDERING VOLUME SOUND SOURCE
    51.
    发明申请
    METHOD AND APPARATUS FOR RENDERING VOLUME SOUND SOURCE 有权

    公开(公告)号：US20220360932A1

    公开(公告)日：2022-11-10

    申请号：US17681429

    申请日：2022-02-25

    Applicant: Electronics and Telecommunications Research Institute

    Inventor： Dae Young JANG ,  Kyeongok KANG ,  Jae-hyoun YOO ,  Yong Ju LEE ,  Tae Jin LEE

    IPC: H04S7/00 ,  H04S3/00

    Abstract: A method and apparatus for rendering a volume sound source are disclosed. The method of rendering a volume sound source may include identifying information about a listener and information about the volume sound source, determining a corresponding area in which a source element is disposed in the volume sound source in consideration of the information about the listener, determining an angle between the listener and the corresponding area based on the information about the listener and the information about the volume sound source, determining a number of source elements disposed in the corresponding area according to the angle, determining a position and a gain of the source element using i) the number of source elements and ii) a distance between the listener and the volume sound source, and rendering the volume sound source according to the position and the gain of the source element.

    DEVICE AND METHOD FOR TRAINING NEURAL NETWORK
    57.
    发明申请
    DEVICE AND METHOD FOR TRAINING NEURAL NETWORK 审中-公开

    公开(公告)号：US20200167659A1

    公开(公告)日：2020-05-28

    申请号：US16696061

    申请日：2019-11-26

    Applicant: Electronics and Telecommunications Research Institute

    Inventor： Yong Hyuk MOON ,  Jun Yong PARK ,  Yong Ju LEE

    IPC: G06N3/08 ,  G06N5/04 ,  G06N20/00 ,  G06F17/16

    Abstract: Provided are a device and method for training a neural network. The method includes generating a candidate solution set by modifying a candidate solution which represents a basic neural network model in a variable-length string form, acquiring first candidate solutions by performing architecture variation-based unsupervised learning with a plurality of candidate solutions selected from the candidate solution set, selecting a neural network model represented by a first candidate solution which satisfies targeted effective performance as a first neural network model, acquiring second candidate solutions by performing selective error propagation-based supervised learning with the first neural network model, and selecting a neural network model represented by a second candidate solution which satisfies the targeted effective performance as a final neural network model.

Patent Agency Ranking