专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明专利

JP2005302011A Method and apparatus for populating electronic forms from scanned documents 有权
标题翻译：用于从扫描文件中计算电子形式的方法和装置
公开(公告)号：JP2005302011A
公开(公告)日：2005-10-27
申请号：JP2005083418
申请日：2005-03-23
申请人： Microsoft Corp , マイクロソフトコーポレーション
发明人： HERLEY CORMAC E , CHELLAPILLA KUMAR H , VIOLA PAUL A , TRAUSTI T KRISTJANSSON
IPC分类号： G06K9/20 , G06F17/00 , G06F17/24 , G06F19/00 , G06K9/03 , G06K9/34 , G06Q10/00
CPC分类号： G06K9/033 , G06F17/243 , G06K9/00449 , G06K9/2054 , G06K2209/01
摘要： PROBLEM TO BE SOLVED: To provide a method and apparatus for populating an electronic form from an electronic image by a computer. SOLUTION: The method and apparatus identify a size, orientation and position of an object within the electronic image, and identify information elements from pixels within the image that corresponds to the object. Fields of the electronic form are displayed to a user along with the identified information elements through a graphical user interface. The information elements are parsed into tagged groups of different information types. At least some of the fields of the electronic form are populated with the tagged groups to produce a populated form. The user is allowed to edit the populated fields through the graphical user interface. COPYRIGHT: (C)2006,JPO&NCIPI
摘要翻译：要解决的问题：提供一种通过计算机从电子图像填充电子表格的方法和装置。该方法和装置识别电子图像内的对象的大小，方向和位置，并且从图像中对应于该对象的像素识别信息元素。通过图形用户界面，电子表格的字段与所标识的信息元素一起显示给用户。信息元素被分析成不同信息类型的标记组。电子表格的至少一些字段填充有标记的组以产生填充形式。允许用户通过图形用户界面编辑填充字段。版权所有（C）2006，JPO＆NCIPI

2. 发明专利

JP2005346707A Low-resolution ocr for document acquired by camera 有权
标题翻译：相机获取的文档的低分辨率OCR
公开(公告)号：JP2005346707A
公开(公告)日：2005-12-15
申请号：JP2005147966
申请日：2005-05-20
申请人： Microsoft Corp , マイクロソフトコーポレーション
发明人： JACOBS CHARLES E , RINKER JAMES R , SIMARD PATRICE Y , VIOLA PAUL A
IPC分类号： G06K9/20 , G06K9/34 , G06K9/46 , G06K9/66 , G06K9/72 , G06N3/00
CPC分类号： G06K9/344 , G06K9/00463 , G06K9/4614 , G06K9/4628 , G06K9/723 , G06K2209/01
摘要： PROBLEM TO BE SOLVED: To provide a global optimization framework for optical character recognition (OCR) of low-resolution photographed documents that combines a binarization-type process, segmentation, and recognition into a single process. SOLUTION: The framework includes a machine learning approach trained on a large amount of data. A convolutional neural network can be employed, to compute a classification function at multiple positions and take gray-level input which eliminates binarization. The framework utilizes preprocessing, layout analysis, character recognition, and word recognition to output high recognition rates. The framework also employs dynamic programming and language models, to arrive at the desired output. COPYRIGHT: (C)2006,JPO&NCIPI
摘要翻译：要解决的问题：为低分辨率拍摄文档的光学字符识别（OCR）提供全局优化框架，将二值化型过程，分割和识别结合到一个过程中。
解决方案：该框架包括针对大量数据训练的机器学习方法。可以采用卷积神经网络来计算多个位置的分类函数，并采用消除二值化的灰度级输入。该框架利用预处理，布局分析，字符识别和字识别来输出高识别率。该框架还采用动态编程和语言模型，以达到所需的输出。版权所有（C）2006，JPO＆NCIPI

3. 发明专利

JP2011008802A Low resolution ocr for camera obtained document 有权
标题翻译：摄像机获取文件的低分辨率OCR
公开(公告)号：JP2011008802A
公开(公告)日：2011-01-13
申请号：JP2010173815
申请日：2010-08-02
申请人： Microsoft Corp , マイクロソフトコーポレーション
发明人： JACOBS CHARLES E , RINKER JAMES R , SIMARD PATRICE Y , VIOLA PAUL A
IPC分类号： G06K9/72 , G06F17/21 , G06K9/20 , G06K9/34 , G06K9/46 , G06K9/66 , G06K9/68 , G06N3/00
CPC分类号： G06K9/344 , G06K9/00463 , G06K9/4614 , G06K9/4628 , G06K9/723 , G06K2209/01
摘要： PROBLEM TO BE SOLVED: To provide a global optimization framework for optical character recognition (OCR) of low-resolution photographed documents, that combines a binarization-type process, segmentation, and recognition into a single process.SOLUTION: The framework includes a machine learning approach trained on a large amount of data. A convolutional neural network can be employed to compute a classification function at multiple positions and take gray-level input which eliminates binarization. The framework utilizes preprocessing, layout analysis, character recognition, and word recognition to output high recognition rates. The framework also employs dynamic programming and language models to arrive at the desired output.
摘要翻译：要解决的问题：为低分辨率拍摄文档的光学字符识别（OCR）提供全局优化框架，将二值化型过程，分割和识别结合到一个过程中。解决方案：框架包括机器学习方法训练了大量的数据。可以采用卷积神经网络来计算多个位置的分类函数，并采用消除二值化的灰度级输入。该框架利用预处理，布局分析，字符识别和字识别来输出高识别率。该框架还采用动态编程和语言模型来达到所需的输出。

4. 发明专利

JP2006073000A Spatial recognition and grouping of text and graphic 有权
标题翻译：空间识别和文本和图形分组
公开(公告)号：JP2006073000A
公开(公告)日：2006-03-16
申请号：JP2005246356
申请日：2005-08-26
申请人： Microsoft Corp , マイクロソフトコーポレーション
发明人： CHELLAPILLA KUMAR H , SHILMAN MICHAEL , VIOLA PAUL A
IPC分类号： G06K9/62 , G06K9/00
CPC分类号： G06K9/726 , G06K9/00402 , G06K9/344 , G06K9/4614 , G06K2209/01
摘要： PROBLEM TO BE SOLVED: To provide a systematic means for recognizing text and/or graphics by using spatial relationships. SOLUTION: By this, a sketched shape is augmented with its symbolic meaning, and numerous features including smart editing, beautification, and interactive simulation of visual languages are made possible. A spatial recognition method obtains a search-based optimization over a large space of possible groupings from simultaneously grouped and recognized sketched shapes. The optimization utilizes a classifier that assigns a class label to a collection of strokes. The overall grouping optimization assumes the properties of the classifier so that if the classifier is scale and rotation invariant, the optimization will be as well. Instances of the present invention employ a variant of AdaBoost to facilitate in recognizing/classifying symbols. Instances of the present invention employ dynamic programming and/or A-star search to perform optimization. COPYRIGHT: (C)2006,JPO&NCIPI
摘要翻译：要解决的问题：提供一种通过使用空间关系识别文本和/或图形的系统手段。
解决方案：通过这一点，草图形状以其符号意义来增强，并且可以实现许多功能，包括智能编辑，美化和视觉语言的交互式模拟。空间识别方法从同时分组和识别的草图形状的可能分组的大空间中获得基于搜索的优化。优化利用了将类标签分配给笔画集合的分类器。总体分组优化假设分类器的属性，以便如果分类器是缩放和旋转不变量，则优化也将同样。本发明的实施例采用AdaBoost的变体来促进识别/分类符号。本发明的实施例采用动态规划和/或A星搜索来执行优化。版权所有（C）2006，JPO＆NCIPI

5. 发明申请

WO2007005937A2 GRAMMATICAL PARSING OF DOCUMENT VISUAL STRUCTURES 审中-公开
标题翻译：文件视觉结构的特征分析
公开(公告)号：WO2007005937A2
公开(公告)日：2007-01-11
申请号：PCT/US2006026140
申请日：2006-06-30
申请人： MICROSOFT CORP
发明人： VIOLA PAUL A , SHILMAN MICHAEL
IPC分类号： G06K9/72 , G06F17/20 , G06K9/34
CPC分类号： G06K9/726 , G06F17/271 , G06K2209/01
摘要： A two-dimensional representation of a document is leveraged to extract a hierarchical structure that facilitates recognition of the document. The visual structure is grammatically parsed utilizing two-dimensional adaptations of statistical parsing algorithms. This allows recognition of layout structures (e.g., columns, authors, titles, footnotes, etc.) and the like such that structural components of the document can be accurately interpreted. Additional techniques can also be employed to facilitate document layout recognition. For example, grammatical parsing techniques that utilize machine learning, parse scoring based on image representations, boosting techniques, and/or "fast features" and the like can be employed to facilitate in document recognition.
摘要翻译：利用文档的二维表示来提取便于识别文档的层次结构。使用统计解析算法的二维适应来语法解析视觉结构。这允许识别布局结构（例如，列，作者，标题，脚注等）等，使得可以准确地解释文档的结构组件。还可以采用附加技术来促进文档布局识别。例如，可以采用利用机器学习，基于图像表示的分析评分，增强技术和/或“快速特征”等的语法解析技术，以促进文档识别。

6. 发明申请

WO2008016392A3 IDENTIFICATION OF PEOPLE USING MULTIPLE TYPES OF INPUT 审中-公开
标题翻译：使用多种类型的输入识别人员
公开(公告)号：WO2008016392A3
公开(公告)日：2008-03-13
申请号：PCT/US2007003715
申请日：2007-02-13
申请人： MICROSOFT CORP
发明人： ZHANG CHA , VIOLA PAUL A , YIN PEI , CUTLER ROSS G , SUN XINDING , RUI YONG
IPC分类号： G01L17/00 , G10L15/24
CPC分类号： G06K9/6256 , G06K9/4614 , G10L25/78 , G10L2021/02166 , H04N7/147 , H04N7/15 , H04N21/42203 , H04N21/4223 , H04N21/4394 , H04N21/44008 , H04N21/44213 , H04N21/4788
摘要： Systems and methods for detecting people or speakers in an automated fashion are disclosed. A pool of features including more than one type of input (like audio input and video input) may be identified and used with a learning algorithm to generate a classifier that identifies people or speakers. The resulting classifier may be evaluated to detect people or speakers.
摘要翻译：公开了以自动方式检测人或扬声器的系统和方法。可以识别包括多于一种类型的输入（如音频输入和视频输入）的特征池并与学习算法一起使用以生成识别人或说话者的分类器。可以评估得到的分类器以检测人或说话者。

7. 发明专利

CA2614177A1 GRAMMATICAL PARSING OF DOCUMENT VISUAL STRUCTURES 未知
公开(公告)号：CA2614177A1
公开(公告)日：2007-01-11
申请号：CA2614177
申请日：2006-06-30
申请人： MICROSOFT CORP
发明人： VIOLA PAUL A , SHILMAN MICHAEL
IPC分类号： G06K9/72 , G06F40/00
摘要： A two-dimensional representation of a document is leveraged to extract a hierarchical structure that facilitates recognition of the document. The visual structure is grammatically parsed utilizing two-dimensional adaptations of statistical parsing algorithms. This allows recognition of layout structures (e.g., columns, authors, titles, footnotes, etc.) and the like such that structural components of the document can be accurately interpreted. Additional techniques can also be employed to facilitate document layout recognition. For example, grammatical parsing techniques that utilize machine learning, parse scoring based on image representations, boosting techniques, and/or “fast features” and the like can be employed to facilitate in document recognition.

8. 发明专利

BRPI0711800A2 未知
公开(公告)号：BRPI0711800A2
公开(公告)日：2011-12-06
申请号：BRPI0711800
申请日：2007-12-13
申请人： MICROSOFT CORP
发明人： ZHANG CHA , VIOLA PAUL A , YIN PEI , GUTLER ROSS G , SUN XINDING , RUI YONG
IPC分类号： G01L17/00 , G10L15/24
摘要： Systems and methods for detecting people or speakers in an automated fashion are disclosed. A pool of features including more than one type of input (like audio input and video input) may be identified and used with a learning algorithm to generate a classifier that identifies people or speakers. The resulting classifier may be evaluated to detect people or speakers.

9. 发明专利

CA2653278C IDENTIFICATION OF PEOPLE USING MULTIPLE TYPES OF INPUT 未知
公开(公告)号：CA2653278C
公开(公告)日：2015-06-23
申请号：CA2653278
申请日：2007-02-13
申请人： MICROSOFT CORP
发明人： ZHANG CHA , VIOLA PAUL A , YIN PEI , CUTLER ROSS G , SUN XINDING , RUI YONG
IPC分类号： A61B5/117 , G10L17/00 , H04N21/80
摘要： Systems and methods for detecting people or speakers in an automated fashion are disclosed. A pool of features including more than one type of input (like audio input and video input) may be identified and used with a learning algorithm to generate a classifier that identifies people or speakers. The resulting classifier may be evaluated to detect people or speakers.

10. 发明专利

ES2390295T3 Identificación de personas usando múltiples tipos de entradas 未知
公开(公告)号：ES2390295T3
公开(公告)日：2012-11-08
申请号：ES07835698
申请日：2007-02-13
申请人： MICROSOFT CORP
发明人： ZHANG CHA , VIOLA PAUL A , YIN PEI , CUTLER ROSS G , SUN XINDING , RUI YONG
IPC分类号： G10L17/00 , G06K9/46 , G06K9/62 , G10L11/02 , G10L21/02 , H04N7/14 , H04N7/15
摘要： Un procedimiento para detección de hablantes que comprende:identificar (310) una serie de características (410, 470) que comprende al menos una característica (420, 430,450) de un primer tipo de entrada y al menos una característica (420, 430, 450) de un segundo tipo de entradadonde el segundo tipo es diferente del primer tipo, en el cual el primer tipo de entrada o el segundo tipo deentrada incluye un entrada de audio (120) y en el cual se calcula una característica para cuantificar algúnelemento del tipo correspondiente de entrada en un tiempo particular; ygenerar (315) un clasificador (785, 855) para la detección de hablantes usando un algoritmo de aprendizaje(745), en el cual el clasificador está constituido por un subconjunto de características, siendo denominado elsubconjunto de características como nodo del clasificador y seleccionado por el algoritmo de aprendizaje,asegurándose también de que los nodos que requieren menos cálculo están situados en el clasificador de talmanera que son evaluados antes que los nodos que requieren más calculo ponderando los nodos mientras segenera el clasificador.

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