SOURCE SEGMENTATION USING Q-CLUSTERING
    1.
    发明申请
    SOURCE SEGMENTATION USING Q-CLUSTERING 审中-公开
    使用Q聚类的源分割

    公开(公告)号:WO2009006265A2

    公开(公告)日:2009-01-08

    申请号:PCT/US2008/068501

    申请日:2008-06-27

    CPC classification number: G10L15/04 G06K9/6226 G10L21/0272 G10L25/78

    Abstract: A system and method that facilitates and effectuates accurate source segmentation of multi-dimensional signals in a computationally efficient manner. By employing Queyranne's algorithm along with a model for combining adjacent multidimensional elements of signal into locally consistent regions, significant improvement in time to identify an optimal segmentation can be achieved. Additional, by saving values computed when executing the algorithm and recalling the values when needed during subsequent iterations of the algorithm provides an additional in algorithm execution speed.

    Abstract translation: 以计算有效的方式促进和实现多维信号的精确源分段的系统和方法。 通过将Queyranne算法与用于将相邻多维信号元素组合为局部一致区域的模型一起使用,可以实现识别最佳分割的时间显着改善。 另外,通过保存在执行算法时计算出的值并在算法的后续迭代期间需要时调用值,可以提供额外的算法执行速度。

    SOURCE SEGMENTATION USING Q-CLUSTERING
    3.
    发明申请
    SOURCE SEGMENTATION USING Q-CLUSTERING 审中-公开
    使用Q-CLUSTERING的资源分类

    公开(公告)号:WO2009006265A3

    公开(公告)日:2009-03-05

    申请号:PCT/US2008068501

    申请日:2008-06-27

    Applicant: MICROSOFT CORP

    CPC classification number: G10L15/04 G06K9/6226 G10L21/0272 G10L25/78

    Abstract: A system and method that facilitates and effectuates accurate source segmentation of multi-dimensional signals in a computationally efficient manner. By employing Queyranne's algorithm along with a model for combining adjacent multidimensional elements of signal into locally consistent regions, significant improvement in time to identify an optimal segmentation can be achieved. Additional, by saving values computed when executing the algorithm and recalling the values when needed during subsequent iterations of the algorithm provides an additional in algorithm execution speed.

    Abstract translation: 一种以计算有效的方式促进并实现多维信号的准确源分割的系统和方法。 通过采用Queyranne的算法以及将相邻的信号多维元素组合成本地一致区域的模型,可以实现显着改善时间以识别最佳分割。 另外,通过保存在执行算法时计算出的值,并且在算法的后续迭代期间需要时调用值提供了额外的算法执行速度。

    EDITING SYSTEM AND EDITING METHOD
    4.
    发明申请
    EDITING SYSTEM AND EDITING METHOD 审中-公开
    编辑系统和编辑方法

    公开(公告)号:WO98045845A1

    公开(公告)日:1998-10-15

    申请号:PCT/JP1998/001596

    申请日:1998-04-07

    Abstract: An editing system provided with an editor (1) which edits video data of a plurality of channels, a disk recorder (2) which is constituted of a disk array having a plurality of disks, and a computer (3). The disk recorder (2) prepares a block map which is used for managing the recording area of each disk so that each sub-block data can be recorded at such a position that the rotation standby time of a head is minimized at the time of accessing to the address of sub-block data recorded on each disk, divides the picture element data of a source video frame into a plurality of sub-block data at every frame based on the block map, and records the sub-block data so that the data can be distributed among a plurality of different disks.

    Abstract translation: 一种具有编辑多个通道的视频数据的编辑器(1)的编辑系统,由具有多个盘的盘阵列构成的盘记录器(2)和计算机(3)。 磁盘记录器(2)准备用于管理每个磁盘的记录区域的块映射,使得每个子块数据可以被记录在这样的位置,即在访问时磁头的旋转待机时间被最小化 记录在每个磁盘上的子块数据的地址,基于块图将每帧的源视频帧的图像元素数据划分为多个子块数据,并记录子块数据,使得 数据可以分布在多个不同的盘之间。

    映像記録装置、映像記録方法
    5.
    发明申请

    公开(公告)号:WO2019016908A1

    公开(公告)日:2019-01-24

    申请号:PCT/JP2017/026186

    申请日:2017-07-20

    Inventor: 桑原 茂樹

    CPC classification number: H04N5/915 H04N5/919 H04N5/92 H04N7/18 H04N21/2312

    Abstract: 従来の技術では、符号化された映像データを2つの記録媒体にフレーム毎に交互に記録する構成になっていたので、1つの記録媒体が故障すると奇数(または偶数)の映像フレームが欠落するためフレームレートが低下するという問題があった。本発明は、データを複数の記録装置へフレーム中の画素データを分散して記録する書き込み制御部と、記録装置が正常か故障かを判断する記録装置故障判断部と、記録装置故障判断部にていずれかの記録装置が故障と判断された場合、正常と判断された記録装置に記録されているデータを用いて故障と判断された記録装置に記録されているデータを補間し、再生データを生成する再生データ生成部とを備える。

    SEPARATE TRACK STORAGE OF TEXTURE AND DEPTH VIEWS FOR MULTIVIEW CODING PLUS DEPTH
    6.
    发明申请
    SEPARATE TRACK STORAGE OF TEXTURE AND DEPTH VIEWS FOR MULTIVIEW CODING PLUS DEPTH 审中-公开
    纹理和深度视图的单独跟踪存储多个编码深度深度

    公开(公告)号:WO2014107378A2

    公开(公告)日:2014-07-10

    申请号:PCT/US2013/077591

    申请日:2013-12-23

    Abstract: Techniques for encapsulating video streams containing multiple coded views in a media file are described. In one example, a method includes parsing a track of multiview video data, wherein the track includes one or more views, including only one of a texture view and a depth view of a particular view. The method further includes parsing a track reference to determine a dependency of the track to a reference track. The track reference types include 'deps' that indicates that the track includes the depth view of the particular view and the reference track includes the texture view of the particular view, 'tref' that indicates that the track depends on the texture view of the particular view which is stored in the referenced track, and 'dref' that indicates that the track depends on the depth view of the particular view which is stored in the referenced track.

    Abstract translation: 描述了将包含多个编码视图的视频流封装在媒体文件中的技术。 在一个示例中,一种方法包括解析多视角视频数据的轨道,其中轨道包括一个或多个视图,其仅包括特定视图的纹理视图和深度视图中的一个。 该方法还包括解析轨道参考以确定轨道对参考轨道的依赖性。 轨道参考类型包括指示轨道包括特定视图的深度视图的“deps”,并且参考轨道包括特定视图的纹理视图,“tref”,其指示轨道取决于特定视图的纹理视图 存储在引用轨道中的视图,以及指示轨道取决于存储在所引用轨道中的特定视图的深度视图的“dref”。

Patent Agency Ranking