专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

11. 发明授权

US07844113B2 Region based push-relabel algorithm for efficient computation of maximum flow 有权
标题翻译：基于区域的推挽算法，用于最大流量的有效计算
公开(公告)号：US07844113B2
公开(公告)日：2010-11-30
申请号：US11685815
申请日：2007-03-14
申请人： Andrew Delong , Yuri Boykov , Daphne Yu
发明人： Andrew Delong , Yuri Boykov , Daphne Yu
IPC分类号： G06K9/34
CPC分类号： G06T7/162 , G06T7/11 , G06T7/12
摘要： A region-based push-relabel formulation is disclosed that removes the requirement that the entire graph should fit into the computer memory and yields an implementation that can reduce the required size and redundancy of accesses to the data memory, thus improving speed performance, while allowing for an efficient parallel processing implementation. The algorithm assigns all vertices that are not part of the sources or sinks with a value of 1. Sinks are assigned with zeros and sources are assigned a label equal to the number of their vertices. The preflow is then pushed from the sources to their neighbors, if any. When the preflow has all reached the boundaries, an adjacent region of the neighboring set is selected and preflow is pushed within this region. When the values of the preflow have been exhausted, region relabeling is done to update the label values. This is repeated within the region until all preflow has exited to the boundary of this region. The operation is then repeated for the neighboring regions that now contain the preflow. Regions which have no preflow may be skipped, thereby realizing a savings in processing resources.
摘要翻译：公开了一种基于区域的推 - 重新标签公式，其消除了整个图形应该适合于计算机存储器的要求，并且产生可以减少对数据存储器的访问所需的大小和冗余度的实现，从而提高速度性能，同时允许用于高效的并行处理实现。算法将不是源或汇的一部分的所有顶点赋值为1。将使用零分配陷阱，并为源分配一个等于其顶点数的标签。然后将预流从源推送到其邻居（如果有的话）。当预流全部达到边界时，选择相邻集合的相邻区域，并且在该区域内推进预流。当预流的值已经耗尽时，进行区域重新标签来更新标签值。这在区域内重复，直到所有预流已经退出到该区域的边界。然后对于现在包含预流的相邻区域重复操作。可以跳过没有预先流动的区域，从而实现了处理资源的节省。

12. 发明申请

US20090295800A1 METHOD FOR DIRECT VOLUMETRIC RENDERING OF DEFORMABLE BRICKED VOLUMES 有权
标题翻译：直径体积增加可变形砖块的方法
公开(公告)号：US20090295800A1
公开(公告)日：2009-12-03
申请号：US12365926
申请日：2009-02-05
申请人： Christoph Vetter , Daphne Yu , Wei Li , Wei Hong
发明人： Christoph Vetter , Daphne Yu , Wei Li , Wei Hong
IPC分类号： G06T17/00
CPC分类号： G06T15/08 , G06T17/005
摘要： A method for rendering a deformable object. The method includes: obtaining a 3D volumetric voxel dataset of a region, such region having therein an object to be rendered; building a tree hierarchical structure for the obtained volumetric dataset, such tree structure blocks as the nodes of a primary tree hierarchy and bricks being those blocks stored as textures in a video memory; augmenting the primary tree hierarchical structure with maximum and minimum values of the data contained within a block; creating a neighborhood tree hierarchy having for each leaf block of the neighborhood tree hierarchy a reference to the neighboring leaf blocks in the neighborhood tree hierarchy as well as references to neighboring bricks in the neighborhood tree hierarchy; updating the information about minimum and maximum in the primary tree hierarchy by saving for each block the minimum and maximum of the neighboring blocks; and rendering the leaf blocks in visibility order.
摘要翻译：一种用于渲染可变形对象的方法。该方法包括：获得区域的3D体积体元数据集，其中具有要渲染的对象的区域; 为所获得的体数据集构建树分层结构，这样的树结构块作为主树层次结构的节点，砖是作为纹理存储在视频存储器中的块; 以包含在块内的数据的最大值和最小值来增加主树分层结构; 创建邻域树层次结构，对于邻域树层次中的每个叶块，对邻域树层次中的相邻叶块的引用以及邻域树层次中的相邻砖的引用; 通过为每个块保存相邻块的最小值和最大值来更新主树层次结构中关于最小和最大值的信息; 并以可见性顺序呈现叶块。

13. 发明授权

US07469073B2 Image-based method for detection and removal of small fragments in segmented three-dimensional volumes 有权
标题翻译：基于图像的方法用于检测和去除分段三维体积中的小片段
公开(公告)号：US07469073B2
公开(公告)日：2008-12-23
申请号：US11110077
申请日：2005-04-20
申请人： Gianluca Paladini , Daphne Yu
发明人： Gianluca Paladini , Daphne Yu
IPC分类号： G06K9/40 , G06K9/00
CPC分类号： G06T15/08 , G06K9/346 , G06K2209/05 , G06T5/005 , G06T7/11 , G06T2200/04 , G06T2207/10081 , G06T2207/30008
摘要： A method for detecting and removing small isolated fragments in a 3D segmented volume is disclosed. The 3D segmented volume is projected onto several 2D images from different viewing directions. Isolated 2D fragments are detected in the 2D images. Corresponding 3D fragments are found in the 3D volume by unprojecting corresponding detected 2D fragment locations. The unprojected detected 2D fragment locations are used as seed points for region growing of isolated 3D fragments. Any of the 3D fragments having a volume size below a user-defined threshold are discarded.
摘要翻译：公开了一种用于检测和去除3D分段体积中的小分离碎片的方法。 3D分割的体积被投影到来自不同观察方向的几个2D图像上。在2D图像中检测到孤立的2D片段。通过取消投影相应的检测到的2D片段位置，在3D体积中找到相应的3D片段。未投影检测到的2D片段位置用作分离的3D片段的区域生长的种子点。任何具有低于用户定义阈值的体积大小的3D片段都被丢弃。

14. 发明申请

US20070286483A1 REGION BASED PUSH-RELABEL ALGORITHM FOR EFFICIENT COMPUTATION OF MAXIMUM FLOW 有权
标题翻译：基于区域的推力相关算法有效计算最大流量
公开(公告)号：US20070286483A1
公开(公告)日：2007-12-13
申请号：US11685815
申请日：2007-03-14
申请人： Andrew Delong , Yuri Boykov , Daphne Yu
发明人： Andrew Delong , Yuri Boykov , Daphne Yu
IPC分类号： G06K9/20
CPC分类号： G06T7/162 , G06T7/11 , G06T7/12
摘要： A region-based push-relabel formulation is disclosed that removes the requirement that the entire graph should fit into the computer memory and yields an implementation that can reduce the required size and redundancy of accesses to the data memory, thus improving speed performance, while allowing for an efficient parallel processing implementation. The algorithm assigns all vertices that are not part of the sources or sinks with a value of 1. Sinks are assigned with zeros and sources are assigned a label equal to the number of their vertices. The preflow is then pushed from the sources to their neighbors, if any. When the preflow has all reached the boundaries, an adjacent region of the neighboring set is selected and preflow is pushed within this region. When the values of the preflow have been exhausted, region relabeling is done to update the label values. This is repeated within the region until all preflow has exited to the boundary of this region. The operation is then repeated for the neighboring regions that now contain the preflow. Regions which have no preflow may be skipped, thereby realizing a savings in processing resources.
摘要翻译：公开了一种基于区域的推 - 重新标签公式，其消除了整个图形应该适合于计算机存储器的要求，并且产生可以减少对数据存储器的访问所需的大小和冗余度的实现，从而提高速度性能，同时允许用于高效的并行处理实现。该算法将不是源或汇的一部分的所有顶点赋值为1. Sinks分配为零，并为源分配一个等于其顶点数的标签。然后将预流从源推送到其邻居（如果有的话）。当预流全部达到边界时，选择相邻集合的相邻区域，并且在该区域内推进预流。当预流的值已经耗尽时，进行区域重新标签来更新标签值。这在区域内重复，直到所有预流已经退出到该区域的边界。然后对于现在包含预流的相邻区域重复操作。可以跳过没有预先流动的区域，从而实现了处理资源的节省。

15. 发明申请

US20070237038A1 Range-Peeling For Punching Masks in Volume Rendering 有权
标题翻译：用于在体积渲染中冲压面罩的范围剥离
公开(公告)号：US20070237038A1
公开(公告)日：2007-10-11
申请号：US11687728
申请日：2007-03-19
申请人： Wei Li , Daphne Yu
发明人： Wei Li , Daphne Yu
IPC分类号： G11B11/00
CPC分类号： G06T15/08
摘要： Rendering with a punching mask is performed without voxelization. A plurality of sub-volumes are identified as a function of the punching mask. The sub-volumes are generally layered in range. Each sub-volume is rendered separately or independently. The renderings from multiple passes or sub-volumes are composited together.
摘要翻译：使用冲孔掩模进行渲染，而不进行体素化。多个子体积被识别为穿孔掩模的函数。子体积通常分层。每个子卷单独或单独渲染。来自多个遍或子卷的渲染合成在一起。

16. 发明授权

US08497861B2 Method for direct volumetric rendering of deformable bricked volumes 有权
标题翻译：直接体积渲染可变形砖体的方法
公开(公告)号：US08497861B2
公开(公告)日：2013-07-30
申请号：US12365926
申请日：2009-02-05
申请人： Christoph Vetter , Daphne Yu , Wei Li , Wei Hong
发明人： Christoph Vetter , Daphne Yu , Wei Li , Wei Hong
IPC分类号： G06T19/00 , G06T17/20 , G06T15/40 , G06T15/08
CPC分类号： G06T15/08 , G06T17/005
摘要： A method for rendering a deformable object. The method includes: obtaining a 3D volumetric voxel dataset of a region, such region having therein an object to be rendered; building a tree hierarchical structure for the obtained volumetric dataset, such tree structure blocks as the nodes of a primary tree hierarchy and bricks being those blocks stored as textures in a video memory; augmenting the primary tree hierarchical structure with maximum and minimum values of the data contained within a block; creating a neighborhood tree hierarchy having for each leaf block of the neighborhood tree hierarchy a reference to the neighboring leaf blocks in the neighborhood tree hierarchy as well as references to neighboring bricks in the neighborhood tree hierarchy; updating the information about minimum and maximum in the primary tree hierarchy by saving for each block the minimum and maximum of the neighboring blocks; and rendering the leaf blocks in visibility order.
摘要翻译：一种用于渲染可变形对象的方法。该方法包括：获得区域的3D体积体元数据集，其中具有要渲染的对象的区域; 为所获得的体数据集构建树分层结构，这样的树结构块作为主树层次结构的节点，砖是作为纹理存储在视频存储器中的块; 以包含在块内的数据的最大值和最小值来增加主树分层结构; 创建邻域树层次结构，对于邻域树层次中的每个叶块，对邻域树层次中的相邻叶块的引用以及邻域树层次中的相邻砖的引用; 通过为每个块保存相邻块的最小值和最大值来更新主树层次结构中关于最小和最大值的信息; 并以可见性顺序呈现叶块。

17. 发明授权

US08170306B2 Automatic partitioning and recognition of human body regions from an arbitrary scan coverage image 失效
标题翻译：从任意扫描覆盖图像自动划分和识别人体区域
公开(公告)号：US08170306B2
公开(公告)日：2012-05-01
申请号：US12107301
申请日：2008-04-22
申请人： Daphne Yu , Marcel Piotraschke , Ludovic Sibille
发明人： Daphne Yu , Marcel Piotraschke , Ludovic Sibille
IPC分类号： G06K9/00
CPC分类号： G06K9/4604 , G06K9/527 , G06K9/726 , G06T7/0012 , G06T2207/10124 , G06T2207/30008
摘要： A recognition pipeline automatically partitions a 3D image of the human body into regions of interest (head, rib cage, pelvis, and legs) and correctly labels each region. The 3D image is projected onto a 2D image using volume rendering. The 3D image may contain the whole body region or any subset. In a learning phase, training datasets are manually partitioned and labeled, and a training database is computed. In a recognition phase, images are partitioned and labeled based on the knowledge from the training database. The recognition phase is computationally efficient and may operate under 2 seconds in current conventional computer systems. The recognition is robust to image variations, and does not require the user to provide any knowledge about the contained regions of interest within the image.
摘要翻译：识别管道将人体的3D图像自动分割成感兴趣的区域（头部，肋骨，骨盆和腿部）并且正确地标记每个区域。使用体积渲染将3D图像投影到2D图像上。 3D图像可以包含整个身体区域或任何子集。在学习阶段，人工分割和标记训练数据集，计算训练数据库。在识别阶段，根据训练数据库的知识对图像进行分区和标记。识别阶段在计算上是有效的，并且可以在现有的传统计算机系统中在2秒以下运行。识别对图像变化是鲁棒的，并且不需要用户提供关于图像内的所包含的感兴趣区域的任何知识。

18. 发明授权

US07808514B2 Multiplanar reformatting visualization with multiple scrolling speeds 有权
标题翻译：多平面重新格式化可视化与多个滚动速度
公开(公告)号：US07808514B2
公开(公告)日：2010-10-05
申请号：US11753663
申请日：2007-05-25
申请人： Daphne Yu , Robert Schneider
发明人： Daphne Yu , Robert Schneider
IPC分类号： G09G5/00
CPC分类号： G06T19/00 , G06F3/0485 , G06T3/0043 , G06T15/08
摘要： Multiplanar reformatting visualization is provided with multiple scrolling speeds. A plurality of buffers is provided. Each buffer stores composites of data from different planes. The amount of compositing to generate a total composite for a visualization may be less since the division into partial composites may reduce addition and subtraction or other compositing operations. One level of partial composite buffers may be used. The total is recomputed for each visualization. In other embodiments, a hierarchal buffer structure is used. For example, a total composite buffer is provided in addition to the partial composite buffers. By re-compositing the partial composites before and after adjustment for scrolling increment from the total composite buffer, the total composite is updated for scrolling. Two or more levels of buffers may be used.
摘要翻译：多平面重新格式化可视化具有多种滚动速度。提供多个缓冲器。每个缓冲存储来自不同平面的数据的合成。用于生成可视化的总复合材料的合成量可能较少，因为对部分复合材料的分割可减少加法和减法或其他合成操作。可以使用一级的部分复合缓冲器。为每个可视化重新计算总数。在其他实施例中，使用层次缓冲结构。例如，除了部分复合缓冲器之外，还提供了总的复合缓冲器。通过在总体复合缓冲区中对组合增量进行调整之前和之后重新合成部分复合材料，总体组合更新为滚动。可以使用两个或多个级别的缓冲器。

19. 发明申请

US20080267471A1 AUTOMATIC PARTITIONING AND RECOGNITION OF HUMAN BODY REGIONS FROM AN ARBITRARY SCAN COVERAGE IMAGE 失效
标题翻译：自动分割和识别人体区域从ARSTIAN SCAN覆盖图像
公开(公告)号：US20080267471A1
公开(公告)日：2008-10-30
申请号：US12107301
申请日：2008-04-22
申请人： Daphne Yu , Marcel Piotraschke , Ludovic Sibille
发明人： Daphne Yu , Marcel Piotraschke , Ludovic Sibille
IPC分类号： G06K9/00
CPC分类号： G06K9/4604 , G06K9/527 , G06K9/726 , G06T7/0012 , G06T2207/10124 , G06T2207/30008
摘要： A recognition pipeline automatically partitions a 3D image of the human body into regions of interest (head, rib cage, pelvis, and legs) and correctly labels each region. The 3D image is projected onto a 2D image using volume rendering. The 3D image may contain the whole body region or any subset. In a learning phase, training datasets are manually partitioned and labeled, and a training database is computed. In a recognition phase, images are partitioned and labeled based on the knowledge from the training database. The recognition phase is computationally efficient and may operate under 2 seconds in current conventional computer systems. The recognition is robust to image variations, and does not require the user to provide any knowledge about the contained regions of interest within the image.
摘要翻译：识别管道将人体的3D图像自动分割成感兴趣的区域（头部，肋骨，骨盆和腿部）并且正确地标记每个区域。使用体积渲染将3D图像投影到2D图像上。 3D图像可以包含整个身体区域或任何子集。在学习阶段，人工分割和标记训练数据集，计算训练数据库。在识别阶段，根据训练数据库的知识对图像进行分区和标记。识别阶段在计算上是有效的，并且可以在现有的传统计算机系统中在2秒以下运行。识别对图像变化是鲁棒的，并且不需要用户提供关于图像内的所包含的感兴趣区域的任何知识。

20. 发明申请

US20050228272A1 System and method for automatically segmenting bones in computed tomography angiography data 失效
标题翻译：计算机断层扫描血管造影数据自动分割骨骼的系统和方法
公开(公告)号：US20050228272A1
公开(公告)日：2005-10-13
申请号：US10915499
申请日：2004-08-09
申请人： Daphne Yu
发明人： Daphne Yu
IPC分类号： A61B5/05 , A61B6/00 , A61B6/03 , G06T5/00 , G06T5/50
CPC分类号： A61B6/504 , A61B6/03 , A61B6/466 , A61B6/481 , G06T7/11 , G06T7/155 , G06T7/174 , G06T2200/04 , G06T2207/10081 , G06T2207/20101 , G06T2207/20224 , G06T2207/30008 , G06T2207/30101
摘要： A system and method for automatically segmenting bone regions from Computed Tomography Angiography (CTA) volume data is disclosed. A locally operated bone detector distinguishes between bone regions and contrast agent filled vessels. A filtering operator removes small noise from the detected bone regions. A dilator expands the filtered detected bone regions to adjacent trabecular bones. A processor applies the dilated detected bone regions to the CTA volume data. A display shows the applied CTA volume data.
摘要翻译：公开了一种用于从计算机断层扫描血管造影（CTA）体数据自动分割骨区域的系统和方法。局部操作的骨检测器区分骨区域和造影剂填充血管。过滤操作员从检测到的骨骼区域中去除小噪声。扩张器将经过检测的骨区域扩展到相邻的小梁骨。处理器将扩大的检测到的骨骼区域应用于CTA体积数据。显示屏显示应用的CTA卷数据。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式