利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

3210 条记录 (耗时 0.044 秒)

    Motion Compensated Picture Boundary Padding
    21.
    发明公开
    Motion Compensated Picture Boundary Padding 审中-公开

    公开(公告)号：US20240314348A1

    公开(公告)日：2024-09-19

    申请号：US18497748

    申请日：2023-10-30

    申请人： TENCENT AMERICA LLC

    发明人： Liang ZHAO ,  Xin ZHAO ,  Shan LIU

    IPC分类号： H04N19/517 ,  H04N19/105 ,  H04N19/137 ,  H04N19/176

    CPC分类号： H04N19/517 ,  H04N19/105 ,  H04N19/137 ,  H04N19/176

    摘要： This disclosure relates generally to video coding and particularly to methods and systems for boundary padding for motion compensation. For example, a motion vector for a motion compensation padded block in a motion compensation padding area of a picture in a video may be determined from scanning a set of neighboring boundary blocks of the picture and/or a set of co-located temporal blocks of the boundary blocks. The motion compensation padded block may be padded according to the motion vector and used for motion compensation of other pictures in the video.

    METHOD AND APPARATUS FOR SECONDARY TRANSFORM WITH ADAPTIVE KERNEL OPTIONS
    22.
    发明公开
    METHOD AND APPARATUS FOR SECONDARY TRANSFORM WITH ADAPTIVE KERNEL OPTIONS 审中-公开

    公开(公告)号：US20240314320A1

    公开(公告)日：2024-09-19

    申请号：US18668876

    申请日：2024-05-20

    申请人： TENCENT AMERICA LLC

    发明人： Madhu PERINGASSERY KRISHNAN ,  Xin ZHAO ,  Shan LIU

    IPC分类号： H04N19/13 ,  H04N19/159 ,  H04N19/176 ,  H04N19/18 ,  H04N19/91

    CPC分类号： H04N19/13 ,  H04N19/159 ,  H04N19/176 ,  H04N19/18 ,  H04N19/91

    摘要： This disclosure relates to secondary transform of video blocks with adaptive kernel options. For example, a method for decoding a video block in a video stream is disclosed. The method may include comprising parsing and processing the video stream to generate: a set of secondary transform coefficients associated with the video block; an intra prediction mode associated with the video block; and a kernel index indicating a secondary transform kernel among a group of secondary transform kernels. The method may further include identifying the group of secondary transform kernels based on the intra prediction mode; and performing an inverse secondary transform of the set of secondary transform coefficients to generate primary transform coefficients of the video block based on the secondary transform kernel among the group of secondary transform kernels identified by the kernel index. The quantity of kernels in the group of secondary transform kernels depends on at least one of: the intra prediction mode associated with the video block; a size of the video block; or a primary transform type associated with the video block.

    METHOD AND APPARATUS FOR DIVIDING PARTIAL SYMMETRY MESH
    23.
    发明公开
    METHOD AND APPARATUS FOR DIVIDING PARTIAL SYMMETRY MESH 审中-公开

    公开(公告)号：US20240312065A1

    公开(公告)日：2024-09-19

    申请号：US18602833

    申请日：2024-03-12

    申请人： TENCENT AMERICA LLC

    发明人： Fang-Yi CHAO ,  Thuong NGUYEN CANH ,  Xiaozhong XU ,  Shan Liu

    IPC分类号： G06T9/00

    CPC分类号： G06T9/001

    摘要： According to one or more embodiments, a method of encoding a mesh includes: determining a global symmetry plane of the mesh that divides the mesh into a first side and a second side; based on the determination that the mesh is partially symmetric: determining each vertex from the plurality of vertices having a symmetry error larger than a first error threshold; performing a clustering process on the plurality of vertices based on the determined symmetry error for each vertex such that each vertex having a symmetry error larger than the first symmetry error threshold is clustered together in one or more clusters; dividing the mesh into a plurality of sub-meshes based on the clustering process; determining whether each sub-mesh is one of fully symmetric, partially symmetric, and asymmetric; and performing symmetry coding on each sub-mesh from the plurality of sub-meshes that is determined to be fully symmetric.

    ADL-UFE: all deep learning unified front-end system
    24.
    发明授权
    ADL-UFE: all deep learning unified front-end system 有权

    公开(公告)号：US12094481B2

    公开(公告)日：2024-09-17

    申请号：US17455497

    申请日：2021-11-18

    申请人： TENCENT AMERICA LLC

    发明人： Yong Xu ,  Meng Yu ,  Shi-Xiong Zhang ,  Dong Yu

    IPC分类号： G10L21/0208 ,  G06N3/044 ,  G06N3/08 ,  G10L21/0216 ,  G10L21/0264 ,  G10L25/30

    CPC分类号： G10L21/0208

    摘要： There is included a method and apparatus comprising computer code for generating enhanced target speech from audio data, performed by a computing device, the method comprising: receiving audio data corresponding to one or more speakers; generating estimated an target speech, an estimated noise, and an estimated echo simultaneously based on the audio data using a jointly trained complex ratio mask; predicting frame-level multi-tap time-frequency (T-F) spatio-temporal-echo filter weights based on the estimated target speech, the estimated noise, and the estimated echo using a trained neural network model; and predicting enhanced target speech based on the frame-level multi-tap T-F spatio-temporal-echo filter weights.

    ANGULAR INTRA PREDICTION
    28.
    发明公开
    ANGULAR INTRA PREDICTION 审中-公开

    公开(公告)号：US20240298022A1

    公开(公告)日：2024-09-05

    申请号：US18215321

    申请日：2023-06-28

    申请人： Tencent America LLC

    发明人： Liang ZHAO ,  Xin ZHAO ,  Jing YE ,  Han GAO ,  Shan LIU

    IPC分类号： H04N19/44 ,  H04N19/105 ,  H04N19/132 ,  H04N19/176

    CPC分类号： H04N19/44 ,  H04N19/105 ,  H04N19/132 ,  H04N19/176

    摘要： Coded information of a current block and a plurality of reference lines of the current block is received. A first offset value is determined in a first mapping table and associated with a first reference line of the plurality of reference lines that is adjacent to the current block. A second offset value is determined in a second mapping table and associated with a second reference line of the plurality of reference lines that is adjacent to the first reference line. A first reference sample for a sample of the current block is determined from the first reference line based on the first offset value. A second reference sample for the sample of the current block is determined from the second reference line based on the second offset value. The sample of the current block is reconstructed based on one of the first reference sample and the second reference sample.

    LOW-POLY MESH GENERATION FOR THREE-DIMENSIONAL MODELS
    29.
    发明公开
    LOW-POLY MESH GENERATION FOR THREE-DIMENSIONAL MODELS 审中-公开

    公开(公告)号：US20240296629A1

    公开(公告)日：2024-09-05

    申请号：US18112445

    申请日：2023-02-21

    申请人： Tencent America LLC

    发明人： Xifeng GAO ,  Zhen CHEN ,  Zherong PAN ,  Kui WU

    IPC分类号： G06T17/20 ,  G06T17/00 ,  G06T19/20

    CPC分类号： G06T17/205 ,  G06T17/005 ,  G06T19/20 ,  G06T2219/2004

    摘要： According to an aspect of the disclosure, a method of generating a low-poly mesh for a three-dimensional (3D) model is provided. In the method, an edge-collapse operation is performed on each of a plurality of edges of an input mesh to generate a simplified mesh in which a number of faces of the input mesh is reduced by the edge-collapse operation. A position of at least one vertex of the simplified mesh is updated to generate an updated mesh in which a distance between the simplified mesh and the input mesh is reduced. The low-poly mesh is generated based on an aligned mesh in which positions of vertices of the updated mesh are optimized by minimizing a shape difference between the simplified mesh and the updated mesh.