利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

684 条记录 (耗时 0.027 秒)

    CONSTRAINED OPTIMIZATION TECHNIQUES FOR GENERATING ENCODING LADDERS FOR VIDEO STREAMING
    2.
    发明公开
    CONSTRAINED OPTIMIZATION TECHNIQUES FOR GENERATING ENCODING LADDERS FOR VIDEO STREAMING 审中-公开

    公开(公告)号：US20240244224A1

    公开(公告)日：2024-07-18

    申请号：US18154680

    申请日：2023-01-13

    申请人： NETFLIX, INC.

    发明人： Sagar Bharadwaj KALASIBAIL SEETHARAM ,  Te-Yuan HUANG ,  Renata Cruz TEIXEIRA ,  Xiaoqing ZHU

    IPC分类号： H04N19/146 ,  H04N19/105

    CPC分类号： H04N19/146 ,  H04N19/105

    摘要： In various embodiments, an encoding ladder application generates encoding ladders that are used to stream media titles. The encoding ladder application generates an objective function based on a ladder configuration and a parameterized objective function. The parameterized objective function approximates a tradeoff between a quality of experience and a cost term associated with a candidate encoding ladder. The encoding ladder application generates constraints based on the ladder configuration and parameterized constraints. The encoding ladder application executes a constrained optimization algorithm on the objective function, the constraints, and encoding point metadata associated with a set of encoded videos to generate a first candidate encoding ladder for a media title.

    MACHINE LEARNING TECHNIQUES FOR VIDEO DOWNSAMPLING
    3.
    发明公开
    MACHINE LEARNING TECHNIQUES FOR VIDEO DOWNSAMPLING 审中-公开

    公开(公告)号：US20240233076A1

    公开(公告)日：2024-07-11

    申请号：US18617162

    申请日：2024-03-26

    申请人： NETFLIX, INC.

    发明人： Li-Heng CHEN ,  Christos G. BAMPIS ,  Zhi LI

    IPC分类号： G06T3/4046 ,  G06N3/084 ,  G06T9/00

    CPC分类号： G06T3/4046 ,  G06N3/084 ,  G06T9/002

    摘要： In various embodiments, a training application trains a convolutional neural network to downsample images in a video encoding pipeline. The convolution neural network includes at least two residual blocks and is associated with a downsampling factor. The training application executes the convolutional neural network on a source image to generate a downsampled image. The training application then executes an upsampling algorithm on the downsampled image to generate a reconstructed image having the same resolution as the source image. The training application computes a reconstruction error based on the reconstructed image and the source image. The training application updates at least one parameter of the convolutional neural network based on the reconstruction error to generate a trained convolutional neural network. Advantageously, the trained convolution neural network can be implemented in a video encoding pipeline to mitigate visual quality reductions typically experienced with conventional video encoding pipelines that implement conventional downsampling techniques.

    Techniques for delivering current media content via content delivery networks
    4.
    发明授权
    Techniques for delivering current media content via content delivery networks 有权

    公开(公告)号：US12003559B1

    公开(公告)日：2024-06-04

    申请号：US18317758

    申请日：2023-05-15

    申请人： NETFLIX, INC.

    发明人： Christopher Alan Newton

    IPC分类号： H04L65/60 ,  G06F8/71 ,  G06F16/955 ,  H04L67/568

    CPC分类号： H04L65/60 ,  G06F8/71 ,  G06F16/955 ,  H04L67/568

    摘要： In various embodiments, a caching application streams segments of a downloadable to a client device. At a first point-in-time, the caching application receives a first request for a first segment of the downloadable from the client device. The caching application computes a cache key based on a request Uniform Resource Locator included in the first request and a version identifier associated with the downloadable. The caching application determines that no segment corresponding to the cache key is stored in a cache. The caching application transmits a second request for the first segment to a different server. Upon receiving a first version of the first segment from the different server, the caching server transmits a response that includes the first version of the first segment to the client device.

    SYSTEMS AND METHODS FOR A LIGHTING ASSEMBLY
    5.
    发明公开
    SYSTEMS AND METHODS FOR A LIGHTING ASSEMBLY 审中-公开

    公开(公告)号：US20240167668A1

    公开(公告)日：2024-05-23

    申请号：US17991795

    申请日：2022-11-21

    申请人： NETFLIX, INC.

    发明人： Paul E. Debevec

    IPC分类号： F21V13/02 ,  F21V3/04 ,  F21V5/00 ,  F21V11/14

    CPC分类号： F21V13/02 ,  F21V3/049 ,  F21V5/007 ,  F21V11/14 ,  F21W2131/406

    摘要： The systems herein include a support structure and multiple light sources mounted to the support structure. The light sources are configured to project light onto a recording stage to light a specified video scene that is to be recorded on the recording stage. These systems also include a perforated layer that includes an arrangement of apertures. The perforated layer has an inward face directed toward the lighting sources and an outward face directed toward the recording stage. The inward face of the perforated layer includes a surface layer that is more reflective than the surface layer of the outward face of the perforated layer. These systems also include a controller that modifies the light emission profile of the light sources, including changing color balance, brightness, time dependence, and/or spatial variation over the light emissive surface of the light sources. Various other apparatuses and recording stage devices are also disclosed.

    Identifying representative frames in video content
    8.
    发明授权
    Identifying representative frames in video content 有权

    公开(公告)号：US11948360B2

    公开(公告)日：2024-04-02

    申请号：US17344752

    申请日：2021-06-10

    申请人： NETFLIX, INC.

    发明人： Shervin Ardeshir Behrostaghi ,  Nagendra K. Kamath

    IPC分类号： G06V20/40 ,  G06V40/16

    CPC分类号： G06V20/47 ,  G06V20/49 ,  G06V40/172

    摘要： One embodiment of the present invention sets forth a technique for selecting a frame of video content that is representative of a media title. The technique includes applying an embedding model to a plurality of faces included in a set of frames of the video content to generate a plurality of face embeddings. The technique also includes aggregating the plurality of face embeddings into a plurality of clusters representing a plurality of characters included in the media title. The technique further includes computing a plurality of prominence scores for the plurality of characters based on one or more attributes of the plurality of clusters, and selecting, from the set of frames, a frame of video content as representative of the media title based on one or more prominence scores for one or more characters included in the frame.

    SYSTEMS AND METHODS FOR CLASSIFYING MUSIC FROM HETEROGENOUS AUDIO SOURCES
    10.
    发明公开
    SYSTEMS AND METHODS FOR CLASSIFYING MUSIC FROM HETEROGENOUS AUDIO SOURCES 审中-公开

    公开(公告)号：US20230409897A1

    公开(公告)日：2023-12-21

    申请号：US17841322

    申请日：2022-06-15

    申请人： Netflix, Inc.

    发明人： Yadong Wang ,  Jeff Kitchener ,  Shilpa Jois Rao

    IPC分类号： G06N3/08 ,  G10H1/00

    CPC分类号： G06N3/08 ,  G10H1/0008 ,  G10H2250/311 ,  G10H2210/036 ,  G10H2210/041

    摘要： The disclosed computer-implemented method may include accessing an audio stream with heterogenous audio content; dividing the audio stream into a plurality of frames; generating a plurality of spectrogram patches, each spectrogram patch within the plurality of spectrogram patches being derived from a frame within the plurality of frames; and providing each spectrogram patch within the plurality of spectrogram patches as input to a convolutional neural network classifier and receiving, as output, a classification of music within a corresponding frame from within the plurality of frames. Various other methods, systems, and computer-readable media are also disclosed.