利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

16 条记录 (耗时 0.021 秒)

    Content collection search with robust content matching
    1.
    发明授权
    Content collection search with robust content matching 有权
    内容集合搜索与强大的内容匹配

    公开(公告)号：US08943090B2

    公开(公告)日：2015-01-27

    申请号：US13621171

    申请日：2012-09-15

    申请人： Arnab S. Dhua ,  Sunil Ramesh ,  Max Delgadillo ,  Raghavan Manmatha

    发明人： Arnab S. Dhua ,  Sunil Ramesh ,  Max Delgadillo ,  Raghavan Manmatha

    IPC分类号： G06F17/30 ,  G06K9/46 ,  G06K9/62

    CPC分类号： G06F17/30256 ,  G06F17/30247 ,  G06F17/30259 ,  G06F17/30542 ,  G06F17/30867 ,  G06K9/46 ,  G06K9/6203

    摘要： Systems and approaches for searching a content collection corresponding to query content are provided. In particular, false positive match rates between the query content and the content collection may be reduced with a minimum content region test and/or a minimum features per scale test. For example, by correlating content descriptors of a content piece in the content collection with query descriptors of the query content, the content piece can be determined to match the query content when a particular region of the content piece and/or a particular region of a query descriptor have a proportionate size meeting or exceeding a specified minimum. Alternatively, or in addition, the false positive match rate between query content and a content piece can be reduced by comparing content descriptors and query descriptors of features at a plurality of scales. A content piece can be determined to match the query content according to descriptor proportion quotas for the plurality of scales.

    摘要翻译： 提供了用于搜索与查询内容相对应的内容集合的系统和方法。 特别地，可以通过最小内容区域测试和/或每个比例测试的最小特征来减少查询内容和内容收集之间的假阳性匹配率。 例如，通过将内容集合中的内容片段的内容描述符与查询内容的查询描述符相关联，可以确定内容片段与内容片段的特定区域和/或内容片段的特定区域 查询描述符的比例大小满足或超过规定的最小值。 或者或另外，通过比较多个尺度的特征的内容描述符和查询描述符，可以减少查询内容与内容片段之间的假正匹配率。 可以根据多个尺度的描述符比例配额来确定内容片段以匹配查询内容。

    CONTENT COLLECTION SEARCH WITH ROBUST CONTENT MATCHING
    2.
    发明申请
    CONTENT COLLECTION SEARCH WITH ROBUST CONTENT MATCHING 有权

    公开(公告)号：US20130254235A1

    公开(公告)日：2013-09-26

    申请号：US13621171

    申请日：2012-09-15

    申请人： Arnab S. Dhua ,  Sunil Ramesh ,  Max Delgadillo ,  Raghavan Manmatha

    发明人： Arnab S. Dhua ,  Sunil Ramesh ,  Max Delgadillo ,  Raghavan Manmatha

    IPC分类号： G06F17/30

    CPC分类号： G06F17/30256 ,  G06F17/30247 ,  G06F17/30259 ,  G06F17/30542 ,  G06F17/30867 ,  G06K9/46 ,  G06K9/6203

    摘要： Systems and approaches for searching a content collection corresponding to query content are provided. In particular, false positive match rates between the query content and the content collection may be reduced with a minimum content region test and/or a minimum features per scale test. For example, by correlating content descriptors of a content piece in the content collection with query descriptors of the query content, the content piece can be determined to match the query content when a particular region of the content piece and/or a particular region of a query descriptor have a proportionate size meeting or exceeding a specified minimum. Alternatively, or in addition, the false positive match rate between query content and a content piece can be reduced by comparing content descriptors and query descriptors of features at a plurality of scales. A content piece can be determined to match the query content according to descriptor proportion quotas for the plurality of scales.

    Scalable tree-based search of content descriptors
    3.
    发明授权
    Scalable tree-based search of content descriptors 有权
    可扩展的基于树的搜索内容描述符

    公开(公告)号：US08352483B1

    公开(公告)日：2013-01-08

    申请号：US12778957

    申请日：2010-05-12

    申请人： Sunil Ramesh ,  Arnab S. Dhua ,  Max Delgadillo ,  Raghavan Manmatha

    发明人： Sunil Ramesh ,  Arnab S. Dhua ,  Max Delgadillo ,  Raghavan Manmatha

    IPC分类号： G06F7/00 ,  G06F17/30

    CPC分类号： G06F17/3002

    摘要： Multiple paths of an index tree may be traversed to discover a set of content descriptors that are match candidates for a set of query descriptors. A size of the set of candidate content descriptors may be optimized, for example, to reduce false positive matching errors, query latencies and/or index tree traversal times, at least in part by determining a number of child nodes to traverse based at least in part on current traverse level and/or traverse neighborhood thresholds. Index trees for large content descriptor sets may be built in resource constrained environments with approximation and/or refining build techniques.

    摘要翻译： 可以遍历索引树的多个路径以发现与一组查询描述符匹配候选的一组内容描述符。 可以优化候选内容描述符集合的大小，例如，至少部分地通过基于至少在...中确定要遍历的子节点的数量来减少假阳性匹配错误，查询延迟和/或索引树遍历时间 部分电流横越电平和/或横越邻域阈值。 用于大内容描述符集的索引树可以在具有近似和/或精炼构建技术的资源约束环境中构建。

    Scalable tree builds for content descriptor search
    4.
    发明授权
    Scalable tree builds for content descriptor search 有权
    可扩展树构建内容描述符搜索

    公开(公告)号：US08756216B1

    公开(公告)日：2014-06-17

    申请号：US12779741

    申请日：2010-05-13

    申请人： Sunil Ramesh ,  Arnab S. Dhua ,  Max Delgadillo ,  Raghavan Manmatha

    发明人： Sunil Ramesh ,  Arnab S. Dhua ,  Max Delgadillo ,  Raghavan Manmatha

    IPC分类号： G06F7/00 ,  G06F17/30

    CPC分类号： G06F17/30946 ,  G06F17/3002

    摘要： Multiple paths of an index tree may be traversed to discover a set of content descriptors that are match candidates for a set of query descriptors. A size of the set of candidate content descriptors may be optimized, for example, to reduce false positive matching errors, query latencies and/or index tree traversal times, at least in part by determining a number of child nodes to traverse based at least in part on current traverse level and/or traverse neighborhood thresholds. Index trees for large content descriptor sets may be built in resource constrained environments with approximation and/or refining build techniques.

    摘要翻译： 可以遍历索引树的多个路径以发现与一组查询描述符匹配候选的一组内容描述符。 可以优化候选内容描述符集合的大小，例如，至少部分地通过基于至少在...中确定要遍历的子节点的数量来减少假阳性匹配错误，查询延迟和/或索引树遍历时间 部分电流横越电平和/或横越邻域阈值。 用于大内容描述符集的索引树可以在具有近似和/或精炼构建技术的资源约束环境中构建。

    Content collection search with robust content matching
    5.
    发明授权
    Content collection search with robust content matching 有权
    内容集合搜索与强大的内容匹配

    公开(公告)号：US08332419B1

    公开(公告)日：2012-12-11

    申请号：US12779254

    申请日：2010-05-13

    申请人： Arnab S. Dhua ,  Sunil Ramesh ,  Max Delgadillo ,  Raghavan Manmatha

    发明人： Arnab S. Dhua ,  Sunil Ramesh ,  Max Delgadillo ,  Raghavan Manmatha

    IPC分类号： G06F7/00 ,  G06F17/00

    CPC分类号： G06F17/30256 ,  G06F17/30247 ,  G06F17/30259 ,  G06F17/30542 ,  G06F17/30867 ,  G06K9/46 ,  G06K9/6203

    摘要： False positive match rates between query content and content in a collection may be reduced with a minimum content region test and/or a minimum features per scale test. The quality of correlations between query descriptors and content descriptors may be improved with a modified sub-region descriptor construction. Content regions associated with detected content features may be partitioned into disjoint sets of sub-regions that cover the content regions, the sub-regions modified so as to at least partially overlap, and descriptor components generated for the modified sub-regions. Matching of feature-sparse content may be improved by adding blurred versions to the collection.

    摘要翻译： 可以通过最小内容区域测试和/或每个比例测试的最小特征来减少查询内容与集合中的内容之间的错误的正匹配率。 可以通过修改的子区域描述符构造来改进查询描述符和内容描述符之间的相关性的质量。 与检测到的内容特征相关联的内容区域可以被划分为覆盖内容区域，修改为至少部分重叠的子区域和为修改的子区域生成的描述符组件的不相交的子区域集合。 通过向集合添加模糊版本，可以改善特征稀疏内容的匹配。

    Method and system for detecting and recognizing text in images
    6.
    发明授权
    Method and system for detecting and recognizing text in images 有权
    检测和识别图像文本的方法和系统

    公开(公告)号：US08009928B1

    公开(公告)日：2011-08-30

    申请号：US12284283

    申请日：2008-09-19

    申请人： Raghavan Manmatha ,  Mark A. Ruzon

    发明人： Raghavan Manmatha ,  Mark A. Ruzon

    IPC分类号： G06K7/10 ,  G06K9/20 ,  G06K9/36 ,  G06K9/34 ,  G06K9/18

    CPC分类号： G06K9/344 ,  G06K9/18 ,  G06K9/325 ,  G06K9/342 ,  G06K9/346 ,  G06K9/38 ,  G06K9/40 ,  G06K9/62 ,  G06K9/6288 ,  G06K9/748 ,  G06K2209/01

    摘要： Various embodiments of the present invention relate to a method, system and computer program product for detecting and recognizing text in the images captured by cameras and scanners. First, a series of image-processing techniques is applied to detect text regions in the image. Subsequently, the detected text regions pass through different processing stages that reduce blurring and the negative effects of variable lighting. This results in the creation of multiple images that are versions of the same text region. Some of these multiple versions are sent to a character-recognition system. The resulting texts from each of the versions of the image sent to the character-recognition system are then combined to a single result, wherein the single result is detected text.

    摘要翻译： 本发明的各种实施例涉及用于检测和识别由相机和扫描仪捕获的图像中的文本的方法，系统和计算机程序产品。 首先，应用一系列图像处理技术来检测图像中的文本区域。 随后，检测到的文本区域通过不同的处理阶段，减少模糊和可变照明的负面影响。 这导致创建多个相同文本区域的图像。 这些多个版本中的一些被发送到字符识别系统。 然后，将发送到字符识别系统的图像的每个版本的结果文本合并为单个结果，其中单个结果是检测到的文本。

    Method and system for representing image patches
    7.
    发明授权
    Method and system for representing image patches 有权
    表示图像补丁的方法和系统

    公开(公告)号：US08406507B2

    公开(公告)日：2013-03-26

    申请号：US12319992

    申请日：2009-01-14

    申请人： Mark A. Ruzon ,  Raghavan Manmatha ,  Donald Tanguay

    发明人： Mark A. Ruzon ,  Raghavan Manmatha ,  Donald Tanguay

    IPC分类号： G06K9/00 ,  G06K9/32

    CPC分类号： G06K9/4671 ,  G06K9/4609 ,  G06K9/6212 ,  G06K9/6232 ,  G06T9/00

    摘要： A method, system and computer program product for representing an image is provided. The image that needs to be represented is represented in the form of a Gaussian pyramid which is a scale-space representation of the image and includes several pyramid images. The feature points in the pyramid images are identified and a specified number of feature points are selected. The orientations of the selected feature points are obtained by using a set of orientation calculating algorithms. A patch is extracted around the feature point in the pyramid images based on the orientations of the feature point and the sampling factor of the pyramid image. The boundary patches in the pyramid images are extracted by padding the pyramid images with extra pixels. The feature vectors of the extracted patches are defined. These feature vectors are normalized so that the components in the feature vectors are less than a threshold.

    摘要翻译： 提供了一种用于表示图像的方法，系统和计算机程序产品。 需要表示的图像以高斯金字塔的形式表示，高斯金字塔是图像的尺度空间表示，并且包括几个金字塔图像。 识别金字塔图像中的特征点，并选择指定数量的特征点。 通过使用一组取向计算算法获得所选特征点的取向。 基于特征点的取向和金字塔图像的采样因子，在金字塔图像的特征点周围提取补丁。 通过用额外的像素填充金字塔图像来提取金字塔图像中的边界补丁。 定义提取的补丁的特征向量。 这些特征向量被归一化，使得特征向量中的分量小于阈值。

    Method and system for representing image patches
    9.
    发明申请
    Method and system for representing image patches 有权
    表示图像补丁的方法和系统

    公开(公告)号：US20100177966A1

    公开(公告)日：2010-07-15

    申请号：US12319992

    申请日：2009-01-14

    申请人： Mark A. Ruzon ,  Raghavan Manmatha ,  Donald Tanguay

    发明人： Mark A. Ruzon ,  Raghavan Manmatha ,  Donald Tanguay

    IPC分类号： G06K9/46

    CPC分类号： G06K9/4671 ,  G06K9/4609 ,  G06K9/6212 ,  G06K9/6232 ,  G06T9/00

    摘要： A method, system and computer program product for representing an image is provided. The image that needs to be represented is represented in the form of a Gaussian pyramid which is a scale-space representation of the image and includes several pyramid images. The feature points in the pyramid images are identified and a specified number of feature points are selected. The orientations of the selected feature points are obtained by using a set of orientation calculating algorithms. A patch is extracted around the feature point in the pyramid images based on the orientations of the feature point and the sampling factor of the pyramid image. The boundary patches in lo the pyramid images are extracted by padding the pyramid images with extra pixels. The feature vectors of the extracted patches are defined. These feature vectors are normalized so that the components in the feature vectors are less than a threshold.

    摘要翻译： 提供了用于表示图像的方法，系统和计算机程序产品。 需要表示的图像以高斯金字塔的形式表示，高斯金字塔是图像的尺度空间表示，并且包括几个金字塔图像。 识别金字塔图像中的特征点，并选择指定数量的特征点。 通过使用一组取向计算算法获得所选特征点的取向。 基于特征点的取向和金字塔图像的采样因子，在金字塔图像的特征点周围提取补丁。 通过用额外的像素填充金字塔图像来提取金字塔图像中的边界斑块。 定义提取的补丁的特征向量。 这些特征向量被归一化，使得特征向量中的分量小于阈值。