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. 发明公开

EP3274918A1 ENHANCING HANDWRITING RECOGNITION USING PRE-FILTER CLASSIFICATION 无效
标题翻译：使用预过滤分类增强手写识别
公开(公告)号：EP3274918A1
公开(公告)日：2018-01-31
申请号：EP16738596.2
申请日：2016-06-24
申请人： Google LLC
发明人： CARBUNE, Victor , DESELAERS, Thomas , KEYSERS, Daniel M.
IPC分类号： G06K9/00 , G06K9/68 , G06K9/22
CPC分类号： G06K9/222 , G06F17/275 , G06K9/00422 , G06K9/6227 , G06K9/6807
摘要： Methods, systems, and devices, including computer programs encoded on a computer storage medium, for improving handwriting detection. In one aspect, a method includes receiving data indicating one or more strokes, determining one or more features of the one or more strokes, determining whether the one or more strokes likely represent a grapheme based at least on one or more of the features, selecting a particular recognition process for processing the data, from among (i) a multi-language recognition process which processes input strokes using multiple recognizers that are each trained to output, for a given set of input strokes, one or more graphemes that are associated with a particular language, and (ii) a single character, universal recognition process which processes input strokes using a universal recognizer that is trained to output, for a given set of input strokes, a single grapheme, and providing the data to the particular recognition process.

2. 发明授权

EP2179381B1 VERFAHREN UND VORRICHTUNG ZUR VERKEHRSZEICHENERKENNUNG 有权转让
标题翻译：方法和装置交通标志识别
公开(公告)号：EP2179381B1
公开(公告)日：2015-06-24
申请号：EP08775189.7
申请日：2008-07-17
申请人： Hella KGaA Hueck & Co.
发明人： BECKER, Lars-Peter
IPC分类号： G06K9/00 , G06K9/68
CPC分类号： G06K9/00818 , G06K9/6807

3. 发明公开

EP2660775A1 BIOMETRIC AUTHENTICATION DEVICE 审中-公开
标题翻译： VORRICHTUNGFÜRBIOMETRISCHE AUTHENTIFIZIERUNG
公开(公告)号：EP2660775A1
公开(公告)日：2013-11-06
申请号：EP10861485.0
申请日：2010-12-27
申请人： Fujitsu Limited
发明人： UNO, Kazuya , ABIKO, Yukihiro
IPC分类号： G06T7/00
CPC分类号： G06K9/0008 , G06K9/00087 , G06K9/00926 , G06K9/6807
摘要： A biometric authentication device, inputs biometric information such as a fingerprint, extracts feature information included in the biometric information, calculates the degree of reliability of the biometric information on the basis of the feature information, obtains a classification of the biometric information on the basis of the feature information, calculates the degree of reliability of the obtained classification, determines whether or not to execute an identification process between the input biometric information and biometric information enrolled in a storage unit on the basis of the degree of reliability of the biometric information and of classification, further determines whether or not a re-input process of the biometric information is needed, executes an identification process for the biometric information when it is determined that the identification process is needed, and issues a re-input instruction for the biometric information when it is determined that the re-input process is needed.
摘要翻译：生物体认证装置输入诸如指纹的生物特征信息，提取生物特征信息中包含的特征信息，根据特征信息计算生物信息的可靠度，基于特征信息获得生物特征信息的分类特征信息，计算所获得的分类的可靠度，基于生物体信息的可靠度确定是否执行输入生物体信息和登记在存储单元中的生物特征信息之间的识别处理，以及进一步确定是否需要生物信息的重新输入处理，当确定需要识别处理时执行生物特征信息的识别处理，并且当生物特征信息的重新输入指令发生时，确定了re-inp 需要进程。

4. 发明公开

EP2558980A2 METHOD AND APPARATUS FOR VISUAL SEARCH STABILITY 有权转让
标题翻译： VERFAHREN UND VORRICHTUNG ZUR STABILISIERUNG VISUELLER SUCHEN
公开(公告)号：EP2558980A2
公开(公告)日：2013-02-20
申请号：EP11768553.7
申请日：2011-04-15
申请人： Nokia Corporation
发明人： GAO, Jiang , JACOB, Matthias , SCHLOTER, Philipp , RUOTTINEN, Pasi
IPC分类号： G06K9/00 , G06F17/30
CPC分类号： G06K9/228 , G06K9/6807 , G06K2009/3291
摘要： Various methods for visual search stability are provided. One example method includes determining a plurality of image matching distances for a captured object depicted in a video frame, where each image matching distance being indicative of a quality of a match between the captured object and a respective object match result. The example method further includes including, in a candidate pool, an indication of the object match results having image matching distances in a candidate region, discarding the object match results having image matching distances in a non-candidate region, and analyzing the object match results with image matching distances in a potential candidate region to include, in the candidate pool, indications of select object match results with image matching distances in the potential candidate region. Similar and related example methods and example apparatuses are also provided.
摘要翻译：提供了用于视觉搜索稳定性的各种方法。一个示例性方法包括为视频帧中描绘的捕获对象确定多个图像匹配距离，其中每个图像匹配距离指示所捕获对象与相应对象匹配结果之间的匹配质量。示例性方法还包括在候选池中包括在候选区域中具有图像匹配距离的对象匹配结果的指示，丢弃在非候选区域中具有图像匹配距离的对象匹配结果，以及分析对象匹配结果具有潜在候选区域中的图像匹配距离，以在候选池中包括在潜在候选区域中具有图像匹配距离的选择对象匹配结果的指示。还提供了类似的和相关的示例性方法和示例设备。

5. 发明公开

EP2283453A1 METHOD AND APPARATUS FOR OBJECT CLASSIFICATION 有权
标题翻译：方法和设备对象分类
公开(公告)号：EP2283453A1
公开(公告)日：2011-02-16
申请号：EP09746039.8
申请日：2009-05-12
申请人： QinetiQ Limited , Advantage West Midlands
发明人： McDONALD, Gregor, John , ELLIS, Jonathan, Stanley , LEVENETS, Douglas, Alan
IPC分类号： G06K9/00 , G06K9/68
CPC分类号： G06K9/00785 , G06K9/6807 , G06K2209/15 , G06K2209/23
摘要： A method of classifying an object, in particular of classifying a vehicle by evaluating the degree of correlation between an image data set corresponding to an image of at least part of the vehicle and each of a plurality of reference data sets each of which corresponds to a reference image of a vehicle comprises the steps of establishing (302) the vehicle's aspect and selecting (304) the plurality of reference data sets from a larger group of such data sets on the basis of the vehicle's aspect. By selecting from the larger group only those reference data sets which correspond to the vehicle's aspect, faster classification is achieved for a given level of processing resources. Alternatively, for a given processing or classification rate, the level of processing resources may be reduced compared to the prior art.

6. 发明公开

EP1587025A3 Biometric information verification 有权
标题翻译：认证的生物特征信息
公开(公告)号：EP1587025A3
公开(公告)日：2009-04-15
申请号：EP04256639.8
申请日：2004-10-27
申请人： FUJITSU LIMITED
发明人： Yamada, Shigefumi, c/o Fujitsu Limited , Shinzaki, Takashi, c/o Fujitsu Limited
IPC分类号： G06K9/00 , G06K9/68
CPC分类号： G06K9/00006 , G06K9/00885 , G06K9/6807
摘要： Every time to-be-verified biometric is input, the apparatus reflects the degree of fluctuation in the to-be-verified biometric information on narrowing down of registered feature information. After to-be-verified biometric information is input through the a biometric information input unit (10), the to-be-verified sub-feature information obtaining unit (30) extracts two or more items of sub-feature information from the to-be-verified biometric information. Based on the pieces of to-be-verified sub-feature information, the to-be-verified sub-feature information assessing unit (40) assesses an acquisition stability of each item of to-be-verifiedsub-featureinformation. Accordingtothe acquisition stability assessment, the registered feature information narrowing unit (50) selects one or more pieces of registered feature information to be subjected to comparison/verification, thereby narrowing down the registered feature information. The apparatus applicable to 1-to- N verification using biometric information such fingerprints, palm prints, iris patterns, facial images, voice patterns, and blood vessel patterns.

7. 发明公开

EP2023268A1 Procédé et dispositif de reconnaissance de position ou de mouvement d'un dispositif ou d'un être vivant 有权转让
标题翻译：的方法及装置或活的设备的位置或运动检测
公开(公告)号：EP2023268A1
公开(公告)日：2009-02-11
申请号：EP08159980.5
申请日：2008-07-09
申请人： Commissariat à l'Energie Atomique
发明人： Bonnet, Stéphane , Godin, Christelle
IPC分类号： G06K9/00 , A61B5/11 , G06K9/22 , G06K9/68
CPC分类号： G06K9/00342 , A61B5/1116 , A61B5/1123 , A61B5/4094 , A61B5/4528 , A61B5/6814 , A61B5/6823 , A61B5/6824 , A61B5/6828 , A61B5/6898 , A61B5/7264 , A61B2560/0418 , A61B2562/0219 , G06F19/00 , G06K9/228 , G06K9/6807
摘要： Ce procédé utilise deux jeux de capteurs pour estimer certaines caractéristiques du mouvement d'un dispositif ou d'un être vivant ou des états, notamment des postures, qu'ils prennent. Un premier jeu, abondant, de capteurs (1) est retiré après une phase d'apprentissage où il a enregistré avec sûreté les états obtenus en exploitant des premières règles de décision. Les mesures d'un second jeu de capteurs (2), beaucoup plus réduit que le premier, sont corrélées aux états atteints pendant l'apprentissage par des secondes règles de décision obtenues automatiquement en alimentant un classificateur. Elles sont ensuite exploitées pour détermine les nouveaux états atteints par le porteur au moyen des seuls seconds capteurs. Les résultats sont bons malgré le petit nombre des seconds capteurs, grâce à la précision des secondes règles de décision.
摘要翻译：该方法包括由两个组传感器应用一组决策规则到位置/运动测量E.G. 陀螺仪（14，15），进行分类的事件E.G. 坐姿。一个知识库，创建在学习阶段，以确定从测量获得的事件分类和特点。另一组规则从基座定义的事件进行分类。一个传感器组被移除并且另一组被保持在笔或生物体即病人。笔/患者的位置/运动期间通过应用后期规则集一认识相分类。因此独立claimsoft被包括为一个位置或运动识别装置包括传感器。

8. 发明公开

EP1247252A4 UNSUPERVISED ADAPTATION AND CLASSIFICATION OF MULTI-SOURCE DATA USING A GENERALIZED GAUSSIAN MIXTURE MODEL 审中-公开
标题翻译：无监督适应和数据从多个来源分类使用广义高斯模型MIX
公开(公告)号：EP1247252A4
公开(公告)日：2007-03-28
申请号：EP00972163
申请日：2000-10-13
申请人： SALK INST , UNIV CARNEGIE MELLON
发明人： LEE TE-WON , LEWICKI MICHAEL , SEJNOWSKI TERRANCE J
IPC分类号： G06F17/18 , G06F17/30 , G06K9/66 , G06K9/68 , G06N3/00 , G10L21/02 , G06K9/62
CPC分类号： G06K9/624 , G06K9/6807 , G10L21/0272
摘要： A computer-implemented method and apparatus that adapts class parameters, classifies data and separates sources configured in one of multiple classes whose parameters are initially unknown. The data set may be generated in a dynamic environment where the sources (100) provide signals are mixed and received by sensors (120), and the mixing parameters change without notice and in an unknown manner. A generalized Gaussian mixture model is used to classify the observed data into two or more mutually exclusive classes whose basis functions may be defined by a range of probability density functions. The class parameters for each of the classes are adapted to a data set in an adaptation algorithm in which class parameters including mixing matrices, bias vectors, and pdf parameters are adapted. Each data vector is assigned to one of the learned mutually exclusive classes. In some embodiments the class parameters may have been previously learned, and the system is used to classify the data and if desired to separate the sources. The adaptation and classification algorithms can be utilized in a wide variety of applications such as speech processing, image processing, medical data processing, satellite data processing, antenna array reception, and information retrieval systems.

9. 发明公开

EP1665086A1 IMPROVING ACCURACY IN SEARCHING DIGITAL INK 审中-公开
标题翻译：提高DIGITAL DYE-搜索结果的准确性
公开(公告)号：EP1665086A1
公开(公告)日：2006-06-07
申请号：EP04761124.9
申请日：2004-08-16
申请人： Silverbrook Research Pty. Limited
发明人： NAPPER, Jonathon L., Silverbrook Research Pty Ltd , LAPSTUN, Paul , SILVERBROOK, Kia
IPC分类号： G06F17/30 , G06K9/00
CPC分类号： G06F17/2765 , G06F17/242 , G06F17/2775 , G06K9/00852 , G06K9/6807 , G10L15/30
摘要： A method and system for improving the accuracy of digital ink (310) searches is disclosed. The method includes receiving a search input query (370) from a user via a user terminal and determining a specialized format of digital ink, by a variety of means, then, based on the determined specialized format of digital ink, a digital ink searching algorithm is selected. A search (380) of a digital ink database (350) can then be performed for a match to the search input query (370) by utilising the digital ink searching algorithm, which is selected from a variety of algorithms so as to improve the accuracy of the search.

10. 发明公开

EP1355240A1 Data analysis method 审中-公开
标题翻译： Verfahren zur Datenanalyse
公开(公告)号：EP1355240A1
公开(公告)日：2003-10-22
申请号：EP02252733.7
申请日：2002-04-18
申请人： BRITISH TELECOMMUNICATIONS public limited company
发明人： The designation of the inventor has not yet been filed
IPC分类号： G06F17/30
CPC分类号： G06F17/30705 , G06K9/6217 , G06K9/6807
摘要： Current classification methods attempt to classify each classification value into a separate class. Consequently, a lot of effort is dedicated to distinguishing between two or more similar classification objects, meaning that supervised learning procedures are slow and produce classifiers that are excessively large. Moreover, the classifiers are often difficult to understand, and take a long time to be generated.
Embodiments of the invention are concerned with reducing the number of classification values that can be used to classify a data item. Relationships between classification values are identified on the basis of attribute values in a set of training data, and those classification values that are determined to be related to one another are subsumed into a single classification group.
An embodiment of the invention is thus concerned with identifying groups of classification values corresponding to a set of data, where each data item in the set is characterised by a plurality of attributes, and each attribute has one of a plurality of attribute values associated therewith. The method comprises the steps of:

(i) selecting an attribute;
(ii) identifying, on the basis of the distribution of attribute values, two classification values that are least similar to one another and allocating a first identified classification value to a first group and a second identified classification value to a second group;
(iii) allocating each unidentified classification value to one of the groups in dependence on correlation between the unidentified classification value and the first and second identified classification values;
(iv) evaluating an association between the first and second groups and the selected attribute;
(v) repeating steps (i) to (iv) for each of at least some of the plurality of attributes
(vi) comparing associations evaluated at step (iv) and selecting first and second groups corresponding to the weakest association;
(vii) for each of the first and second groups repeating steps (i) to (vi) for the classification values therein, until the association evaluated at step (iv) falls below a predetermined threshold value.

Essentially classification groups are repeatedly analysed with respect to a range of attributes so as to identify all possible groupings of classification values. For example, classification values Daily Mail, Daily Express, The Times, The Guardian, Vogue, New Scientist, Economist, Cosmopolitan, FHM, House and Garden are analysed with respect to a selection of attributes (e.g. sex, age, occupation etc.). Assuming that the analysis identifies the classification values as falling within two classification groups: [Daily Mail, Daily Express, Cosmopolitan, FHM] and [The Times, The Guardian, Vogue, New Scientist, Economist, House and Garden], each of these groups is then analysed with respect to the same, or a different, selection of attributes. This second round of analysis may identify further clusters of classification values - e.g. the analysis could show that the classification values in the latter group are clustered into two distinct groups: [ House and Garden, Vogue ] and [ The Times, The Guardian, New Scientist, Economist ]. After each round of analysis an association between the groups and attributes is measured and is compared with a threshold; this comparison identifies whether or not the groups are sufficiently different as to justify splitting up the classification values into groups.
摘要翻译：当前的分类方法尝试将每个分类值分类到一个单独的类中。因此，大量的努力专门用于区分两个或更多个类似的分类对象，这意味着监督学习过程缓慢并且产生过大的分类器。此外，分类器往往难以理解，需要很长时间才能产生。本发明的实施例涉及减少可用于对数据项进行分类的分类值的数量。基于一组训练数据中的属性值来识别分类值之间的关系，并且被确定为彼此相关的那些分类值被归入单个分类组。因此，本发明的实施例涉及识别与一组数据相对应的分类值组，其中集合中的每个数据项由多个属性表征，并且每个属性具有多个属性中的一个与之相关联的值。该方法包括以下步骤：（i）选择属性; （ii）基于属性值的分布来识别彼此最不相似的两个分类值，并将第一识别的分类值分配给第一组，将第二识别分类值分配给第二组; （iii）根据未识别的分类值与第一和第二识别的分类值之间的相关性，将每个未识别的分类值分配给组中的一个; （iv）评估第一组和第二组之间的关联以及所选择的属性; （v）针对所述多个属性中的至少一些属性（vi）中的每一个重复步骤（i）至（iv），以比较在步骤（iv）处评估的关联并选择对应于最弱关联的第一和第二组; （vii）对于第一组和第二组中的每一个，对于其中的分类值重复步骤（i）至（vi），直到在步骤（iv）评估的关联低于预定阈值。相对于一系列属性重复分析基本分类组，以便识别所有可能的分类值分组。例如，分类价值观分析了每一个“每日邮报”，“每日快报”，“时代报”，“卫报”，“时尚”，“新科学家”，“经济学家”，“大都会”，“家庭与家庭” 。假设分析确定分类值属于两个分类组：Ä日报，每日快报，大都会，FHMÜ和ÄTheTimes，“卫报”，“时尚”，“新科学家”，“经济学家”，“House and Garden” 尊重相同或不同的属性选择。第二轮分析可以识别进一步的分类值簇 - 例如。分析可以表明，后一组的分类值分为两个不同的组：ÄHouseand Garden，VogueÜ和ÄTheTimes，The Guardian，New Scientist，EconomistÜ。在每轮分析之后，测量组和属性之间的关联并将其与阈值进行比较; 该比较确定组是否有足够的不同，以证明将分类值分成组。

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