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

JP2017084058A 画像処理装置、画像処理方法及び画像処理プログラム审中-公开
公开(公告)号：JP2017084058A
公开(公告)日：2017-05-18
申请号：JP2015210875
申请日：2015-10-27
申请人：株式会社東芝 , Toshiba Corp
发明人： IMOTO KAZUNORI , TOUCHI YOJIRO , SUZUKI KAORU , YAMAGUCHI OSAMU
IPC分类号： G06K9/20 , G06K9/03
CPC分类号： G06K9/2081 , G06K9/342 , G06K2209/01
摘要：【課題】簡単な操作で効率的に文字を読み取り可能な画像処理装置、画像処理方法及び画像処理プログラムを提供する。【解決手段】実施形態によれば、取得部と、処理部と、を備えた画像処理装置が提供される。前記取得部は、複数の文字列を含む画像を取得する。前記処理部は、検出動作と、受取動作と、抽出動作と、生成動作と、を実施する。前記検出動作は、前記画像から前記複数の文字列に関する複数の画像領域を検出することを含む。前記受取動作は、前記画像内の座標に関する座標情報の入力を受け取ることを含む。前記抽出動作は、前記座標情報により指定される指定領域を、前記複数の画像領域の中から抽出することを含む。前記生成動作は、前記座標情報に基づいて、前記指定領域の数及びサイズの少なくともいずれかを修正した修正領域を生成することを含む。【選択図】図1

2. 发明专利

JP2011217179A Apparatus and method for canceling echo 有权
标题翻译：装置和方法取消ECHO
公开(公告)号：JP2011217179A
公开(公告)日：2011-10-27
申请号：JP2010084318
申请日：2010-03-31
申请人： Toshiba Corp , 株式会社東芝
发明人： SUZUKI KAORU , YAMAMOTO KOICHI , AMADA KO
IPC分类号： H04B3/23 , H04M1/60
CPC分类号： H04M9/082
摘要： PROBLEM TO BE SOLVED: To provide an apparatus and method for canceling an echo, capable of accurately suppressing an acoustic echo.SOLUTION: A filter is configured to generate a third acoustic signal by convoluting a first acoustic signal with a predetermined tap coefficient. A subtraction unit is configured to generate a fourth acoustic signal by subtracting the third acoustic signal from a second acoustic signal. An estimation unit is configured to determine whether the sound volume of the first acoustic signal is below a predetermined threshold, and to set the sound volume of the second acoustic signal as a first value when the determination is affirmative. A correction unit is configured to correct the tap coefficient by using a determiner configured to determine a step size used for correcting the tap coefficient based on the first value, and the determined step size.
摘要翻译：要解决的问题：提供一种抵消回声的装置和方法，能够精确地抑制声学回声。解决方案：滤波器被配置为通过以预定的抽头系数卷积第一声信号来产生第三声信号。减法单元被配置为通过从第二声信号中减去第三声信号来产生第四声信号。估计单元被配置为确定第一声信号的音量是否低于预定阈值，并且当确定为肯定时，将第二声信号的音量设置为第一值。校正单元被配置为通过使用被配置为基于第一值确定用于校正抽头系数的步长的确定器以及所确定的步长来校正抽头系数。

3. 发明专利

JP2011203700A Sound discrimination device 有权
标题翻译：声音辨别装置
公开(公告)号：JP2011203700A
公开(公告)日：2011-10-13
申请号：JP2010073700
申请日：2010-03-26
申请人： Toshiba Corp , 株式会社東芝
发明人： YAMAMOTO KOICHI , SUZUKI KAORU , AMADA KO
IPC分类号： G10L11/00 , G10L15/04 , G10L15/20 , G10L21/02
CPC分类号： G10L25/78 , G10L21/0208
摘要： PROBLEM TO BE SOLVED: To suppress a remaining echo component included in a feature amount.SOLUTION: A sound discriminating apparatus 100 includes: a first acoustic signal analyzing unit 103 configured to analyze a frequency spectrum of a first acoustic signal; and a feature extracting unit 101 configured to remove a frequency spectrum of the first acoustic signal from a third acoustic signal, which is obtained by suppressing an echo component of the first acoustic signal contained in a second acoustic signal, so as to extract the feature amount of a frequency spectrum of the third acoustic signal.
摘要翻译：要解决的问题：抑制特征量中包含的剩余回声成分。解决方案：声音识别装置100包括：第一声音信号分析单元103，被配置为分析第一声信号的频谱; 以及特征提取单元101，被配置为从通过抑制包含在第二声信号中的第一声信号的回波分量获得的第三声信号中去除第一声信号的频谱，以便提取特征量的第三声信号的频谱。

4. 发明专利

JP2010032763A Device and method for supporting study of language 有权
标题翻译：用于支持语言研究的设备和方法
公开(公告)号：JP2010032763A
公开(公告)日：2010-02-12
申请号：JP2008194692
申请日：2008-07-29
申请人： Toshiba Corp , 株式会社東芝
发明人： CHINO TETSURO , SHIBAZAKI YASUYO , O RAN , SATO NAOKO , KIKUCHI MASAAKI , SUZUKI KAORU
IPC分类号： G09B19/06 , G09B5/06
摘要： PROBLEM TO BE SOLVED: To provide a language study supporting device enabling a user to efficiently learn expressions in actual communication. SOLUTION: The language study supporting device includes: a language knowledge storage section 121 storing language knowledge to be used in natural language analysis; a description storage section 122 which stores the language knowledge in association with description information showing the description of the language knowledge; an input section 101 used for inputting an input character string expressing either the user's or the other person's utterance; an analysis section 102 which analyzes the input character string on the basis of the language knowledge and identifies the language knowledge used for the analysis; a speaker determining section 103 determining whether the input character string corresponds to the user's or the other person's utterance; an description acquisition section 105 which acquires, when the input character string corresponds to the other person's utterance, the description information associated with the identified language knowledge from the description storage section; and an output section 106 for outputting the acquired description information to the user. COPYRIGHT: (C)2010,JPO&INPIT
摘要翻译：要解决的问题：提供语言学习支持装置，使得用户能够在实际通信中有效地学习表达。语言学习辅助设备包括：语言知识存储部分121，用于存储自然语言分析中使用的语言知识; 描述存储部分122，其与表示语言知识的描述的描述信息相关联地存储语言知识; 用于输入表示用户或其他人的话语的输入字符串的输入部分101; 分析部分102，其基于语言知识分析输入字符串，并识别用于分析的语言知识; 扬声器确定部分103，确定输入的字符串是否对应于用户或其他人的话语; 描述获取部分105，当输入字符串对应于另一个人的话语时，从描述存储部分获取与所识别的语言知识相关联的描述信息; 以及用于将所获取的描述信息输出给用户的输出部分106。版权所有（C）2010，JPO＆INPIT

5. 发明专利

JP2009229899A Device and method for voice recognition 审中-公开
标题翻译：用于语音识别的设备和方法
公开(公告)号：JP2009229899A
公开(公告)日：2009-10-08
申请号：JP2008076275
申请日：2008-03-24
申请人： Toshiba Corp , 株式会社東芝
发明人： YAMAMOTO DAISUKE , SUGIYAMA HIROSHI , KOGA TOSHIYUKI , SUZUKI KAORU
IPC分类号： G10L15/28 , G10L15/00
CPC分类号： G10L15/20 , G10L21/0208 , G10L25/78
摘要： PROBLEM TO BE SOLVED: To reduce recognition error due to noise, when operating a device by using voice recognition.
SOLUTION: A voice recognition device for recognizing input voice from a microphone 1 comprises: a voice obtaining section 2 for obtaining the input voice from the microphone 1; a voice period detection section 3 for detecting uttering start time and uttering end time based on the input voice; a voice identification section 4 for identifying the input voice; a distance measuring sensor 11 for emitting infra-red ray for measuring a distance between the microphone 1 and a user 10, which may become an electric noise source; and a noise generation suppressing means for suppressing noise generation from the electric noise source, which reaches the microphone 1, by temporarily stopping operation of the distance measuring sensor 11, in the voice period from the uttering start time to the uttering end time, which is detected by the voice period detection section 3. A gain of the microphone 1 is adjusted according to the distance of the microphone 1 and the user 10, which is measured by the distance measuring sensor 11.
COPYRIGHT: (C)2010,JPO&INPIT
摘要翻译：要解决的问题：通过使用语音识别来减少由噪声引起的识别误差。解决方案：用于从麦克风1识别输入语音的语音识别装置包括：语音获取部分2，用于从麦克风1获得输入的声音; 声音周期检测部分3，用于基于输入的声音来检测发声开始时间和发声结束时间; 用于识别输入语音的语音识别部分4; 距离测量传感器11，用于发射用于测量麦克风1与可能成为电噪声源的用户10之间的距离的红外线; 以及噪声生成抑制装置，用于在从发出开始时间到发出结束时间的语音周期中通过临时停止距离测量传感器11的操作来抑制到达麦克风1的电噪声源的噪声产生，这是从根据由距离测量传感器11测量的麦克风1和用户10的距离来调节麦克风1的增益。（C）2010， JPO和INPIT

6. 发明专利

JP2009192942A Voice interaction apparatus and support method 有权
标题翻译：语音交互设备和支持方法
公开(公告)号：JP2009192942A
公开(公告)日：2009-08-27
申请号：JP2008035126
申请日：2008-02-15
申请人： Toshiba Corp , 株式会社東芝
发明人： SUGIYAMA HIROSHI , SUZUKI KAORU , YAMAMOTO DAISUKE , KOGA TOSHIYUKI
IPC分类号： G10L15/00 , G10L15/22 , H04N5/00
CPC分类号： G10L15/22 , G10L15/20
摘要： PROBLEM TO BE SOLVED: To provide an interactive robot capable of estimating an area suitable for user's uttering in order to improve voice recognition accuracy. SOLUTION: The voice interactive robot 100 capable of performing voice recognition performs processing corresponding to a word recognized by a voice recognition section 302. A recommended distance range for satisfying a necessary S/N ratio is estimated from voice characteristic of a user 201 which has been stored in a user voice intensity DB 306, and it is output to a recommended distance range presentation section 309. The recommended distance range presentation section 309 presents a recommended distance range which is a distance range suitable for voice of the interactive robot 100 and the user 201 which are notified by a recommended distance range estimating section 307. Thereby, the interactive robot can be provided which can perform support in such a manner that the user may utter from the recommended distance range suitable for uttering which satisfies the voice recognition accuracy. COPYRIGHT: (C)2009,JPO&INPIT
摘要翻译：要解决的问题：提供一种能够估计适于用户发声的区域的交互式机器人，以便提高语音识别精度。解决方案：能够执行语音识别的语音交互式机器人100执行与由语音识别部分302识别的单词相对应的处理。用于满足必要S / N比的推荐距离范围是根据用户201的语音特征其被存储在用户声音强度DB 306中，并且被输出到推荐的距离范围呈现部分309.推荐的距离范围呈现部分309呈现推荐的距离范围，其是适合于交互式机器人100的语音的距离范围以及由推荐距离范围估计部分307通知的用户201.由此，可以提供可以执行支持的交互式机器人，使得用户可以从适合于发声的推荐距离范围发出，其满足语音识别准确性。版权所有（C）2009，JPO＆INPIT

7. 发明专利

JP2009080576A Retrieving apparatus, method, and program 审中-公开
标题翻译：检索设备，方法和程序
公开(公告)号：JP2009080576A
公开(公告)日：2009-04-16
申请号：JP2007247992
申请日：2007-09-25
申请人： Toshiba Corp , 株式会社東芝
发明人： DOI MIWAKO , SUZUKI KAORU , KOGA TOSHIYUKI , YAMAMOTO KOICHI
IPC分类号： G06F17/30 , H04N7/173 , H04N21/432
CPC分类号： G06F17/30817
摘要： PROBLEM TO BE SOLVED: To provide a retrieving apparatus for efficiently retrieving a content material related to the names and aliases represented by a keyword even when one keyword is inputted to be retrieved from a user, and to provide a retrieving method, and a retrieving program. SOLUTION: A word corresponding to a keyword to be retrieved and a word associated with the word are read from a word dictionary in which words representing proper names and aliases are registered in association with a relation attribute indicating relation among the words. The keyword is expanded by reading the words, and content related to any one of the read words is retrieved. COPYRIGHT: (C)2009,JPO&INPIT
摘要翻译：要解决的问题：提供一种检索装置，用于有效地检索与由关键字表示的名称和别名相关的内容素材，即使输入一个关键字以从用户检索，并提供检索方法，以及检索程序。解决方案：从一个单词字典中读取与要检索的关键字对应的单词和与单词相关的单词，其中表示专有名称和别名的单词与指示单词之间关系的关系属性相关联地登记。通过阅读单词来扩展关键字，并检索与任何一个读取单词相关的内容。版权所有（C）2009，JPO＆INPIT

8. 发明专利

JP2008242504A User interface device, display method and display program 审中-公开
标题翻译：用户界面设备，显示方法和显示程序
公开(公告)号：JP2008242504A
公开(公告)日：2008-10-09
申请号：JP2007077582
申请日：2007-03-23
申请人： Toshiba Corp , 株式会社東芝
发明人： DOI MIWAKO , SUZUKI KAORU , KOGA TOSHIYUKI , YAMAMOTO KOICHI
IPC分类号： G06F3/048 , G06F3/0482
CPC分类号： H04N5/44543 , G11B27/329 , H04N5/76 , H04N5/765 , H04N21/4147 , H04N21/4334 , H04N21/4335 , H04N21/44222 , H04N21/4622 , H04N21/4828 , H04N21/84
摘要： PROBLEM TO BE SOLVED: To provide a user interface device, method and program, improving operability related to instruction input to a content material. SOLUTION: A property related to a reproduction state or a temporal property of content information with information related to the content material and/or the content material as an operation candidate content is distinguished, the operation candidate content is list-displayed for each kind of the property, and processing contents corresponding to the property of an operation target content is displayed as a processing candidate. COPYRIGHT: (C)2009,JPO&INPIT
摘要翻译：要解决的问题：提供一种用户界面设备，方法和程序，提高与内容资料的指令输入相关的可操作性。解决方案：区分与具有与作为操作候选内容的内容资料和/或内容素材有关的信息的内容信息的再现状态或时间属性的属性，对每个操作候选内容进行列表显示作为属性的种类和与操作对象内容的属性对应的处理内容被显示为处理候选。版权所有（C）2009，JPO＆INPIT

9. 发明专利

JP2006087827A Diagnostic imaging apparatus and image processing system, program and method 审中-公开
标题翻译：诊断成像设备和图像处理系统，程序和方法
公开(公告)号：JP2006087827A
公开(公告)日：2006-04-06
申请号：JP2004280207
申请日：2004-09-27
申请人： Toshiba Corp , Toshiba Medical Systems Corp , 東芝メディカルシステムズ株式会社 , 株式会社東芝
发明人： SUZUKI KAORU
IPC分类号： A61B8/00 , A61B5/055 , A61B6/03 , G06T1/00
摘要： PROBLEM TO BE SOLVED: To provide a diagnostic imaging apparatus capable of appropriately extracting image data in a three-dimensional region even if it is inappropriate to perform a linear interpolation perpendicular to a slice image when extracting the image data in the three-dimensional region.
SOLUTION: This diagnostic imaging apparatus is provided with a means 43a of detecting a position of the gravity center of a reference ROI (Region Of Interest) set in two-dimensional tomographic data, an ROI gravity center position accordance means 43b uniformly moving the two-dimensional tomographic data parallel so that the position of the gravity center of the reference ROI is positioned on a straight line perpendicular to the slice, a perpendicular linear interpolation means 43e setting the three-dimensional region by performing the linear interpolation to the two-dimensional tomographic data after the parallel movement and setting a part included in the set three-dimensional region as a region of interest, and an ROI gravity center position return means 43f moving the two-dimensional tomographic data parallel so that the position of the gravity center of the two-dimensional tomographic data after the parallel movement is returned to the position of the gravity center before the parallel movement.
COPYRIGHT: (C)2006,JPO&NCIPI
摘要翻译：要解决的问题：提供一种能够在三维区域中适当地提取图像数据的诊断成像装置，即使当在三维区域中提取图像数据时不适合执行垂直于切片图像的线性插值时，三维区域。解决方案：该诊断成像装置设置有检测设置在二维断层数据中的参考ROI（感兴趣区域）的重心位置的装置43a，均匀移动的ROI重心位置对齐装置43b 二维断层数据平行，使得参考ROI的重心位置位于与切片垂直的直线上，垂直线性插值装置43e通过对两维进行线性插值来设定三维区域平行移动后的三维断层摄影数据和将包含在所设置的三维区域中的部分设置为感兴趣区域，以及ROI重心位置返回装置43f平行移动二维断层数据，使得重力位置平行运动后的二维断层数据的中心返回到重心之前的位置平行运动。版权所有（C）2006，JPO＆NCIPI

10. 发明专利

JP2004028990A Image processor and method for the same 有权
公开(公告)号：JP2004028990A
公开(公告)日：2004-01-29
申请号：JP2003062529
申请日：2003-03-07
申请人： Toshiba Corp , 株式会社東芝
发明人： WADA TAKASHI , SUZUKI KAORU , NAKAMURA TATSURO , MATSUDA NATSUKO
IPC分类号： G01B11/00 , G06T1/00
摘要： PROBLEM TO BE SOLVED: To provide an image processor which surely calculates the spatial position on an object to be observed, even when a spike noise occurs.
SOLUTION: The image processor is composed of a light pattern projection part 501, such as a laser light source or the like, for compulsorily generating a light pattern on the surface of the object, imaging parts 502 and 505 which are a first to n-th image input means for imaging the generated pattern, peak detection parts 503 and 506 for calculating the brightness distribution of the taken image and for detecting the peak value of the brightness therefrom, position calculation parts 504 and 507 for calculating spatial coordinates of the sampled point on the object from the point detected by the peak detection parts, a score calculation part 508 for classifying the points on the object calculated by the first to the n-th imaging parts 502 and 505 into clusters within ε-range and for determining the score of the cluster by the number of the points in the cluster, and an output part 509 for outputting, as spatial position information, the representative point of the cluster obtaining the highest score at the score calculation part, and the spatial position is calculated.
COPYRIGHT: (C)2004,JPO

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