会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 82. 发明专利
    • 画像処理システム及び画像処理方法
    • 一种图像处理系统和图像处理方法
    • JPWO2013089261A1
    • 2015-04-27
    • JP2013549343
    • 2012-12-10
    • 日本電気株式会社
    • 剛志 柴田剛志 柴田彰彦 池谷彰彦 池谷仙田 修司修司 仙田
    • G06T3/00H04N1/387
    • G06T5/003G06K9/6256G06K9/66G06T5/50G06T9/008G06T2207/20024G06T2207/20081G06T2207/20182G06T2207/20201H04N19/97
    • 本発明は、低解像入力画像に対応して得られる復元画像を、本来の画像に十分近くする画像処理装置を提供する。その画像処理装置は、所定の画像を劣化させた劣化画像の劣化パッチと、その所定の画像のパッチである復元パッチとを対応付けたデータを格納する辞書を利用し、入力画像を分割して生成した複数の入力パッチと劣化パッチとの類似度として、それらの劣化パッチ及び入力パッチそれぞれの形状を連続的な重みであるパッチ重みで再設定した、重み付き劣化パッチ及び重み付き入力パッチ間の類似度である重み付き類似度を算出する類似度算出手段と、その重み付き類似度に基づいて、入力パッチ毎に復元パッチを選択する選択手段と、その復元パッチを合成して復元画像を生成する合成手段とを備える。
    • 本发明是得到了对应于低分辨率输入图像,提供一种图像处理装置是足够接近原始图像的恢复图像。 所述的图像处理装置利用的字典用于存储退化图像以降解预定的图像的劣化的补丁,修复补片相关联的数据是预定的图像的补片,通过将所述输入图像 所生成的多个输入补丁和降解贴片的相似性,是重新配置补丁加权,加权恶化补丁和加权输入补丁它们的降解贴片和输入接插各自的形状和重量不断之间 产生相似性计算装置,用于计算加权相似度是基于加权的类似的选择相似性装置,用于选择针对每一输入补丁还原补丁,恢复的图像恢复该补丁合成并 和一合成装置,用于。
    • 85. 发明专利
    • Object detection method and object detection device
    • 对象检测方法和对象检测装置
    • JP2013210967A
    • 2013-10-10
    • JP2012082379
    • 2012-03-30
    • Canon Incキヤノン株式会社
    • TSUKAMOTO KENJITORII HIROSHIMATSUGI MASAKAZU
    • G06T7/00G06T1/00
    • G06K9/00362G06K9/3241G06K9/66G06K9/6857G06T3/40
    • PROBLEM TO BE SOLVED: To provide a method for performing detection processing fast while maintaining accuracy.SOLUTION: An image including a target object is captured and the image is magnified by a plurality of different magnifications to generate a layered image. Based on the layered image, a region in at least a part of the target object is detected as a first detection region and based on the first detection region, a first specific parts region is estimated. Based on the first specific parts region and learning data which are learnt beforehand, a layer of the layered image is restricted. In the layered image of the restricted layer, a region in at least a part of the target object is detected as a second detection region and based on the second detection region, a second specific parts region is estimated. The first specific parts region and the second specific parts region are integrated and determined as a specific parts region of the target object.
    • 要解决的问题:提供一种在保持精度的同时快速进行检测处理的方法。解决方案:捕获包括目标对象的图像,并且通过多个不同的放大倍数放大图像以生成分层图像。 基于分层图像,检测目标对象的至少一部分中的区域作为第一检测区域,并且基于第一检测区域,估计第一特定部位区域。 基于预先学习的第一特定部位区域和学习数据,分层图像的层被限制。 在限制层的分层图像中,目标对象的至少一部分中的区域被检测为第二检测区域,并且基于第二检测区域,估计第二特定区域。 第一特定部位区域和第二特定部位区域被集成并确定为目标物体的特定部位区域。
    • 86. 发明专利
    • Image processing device, image processing method and program
    • 图像处理设备,图像处理方法和程序
    • JP2013134628A
    • 2013-07-08
    • JP2011284566
    • 2011-12-26
    • Fujitsu Ltd富士通株式会社
    • NAKAO MANABU
    • G06T7/00
    • G06K9/4642G06K9/6212G06K9/6215G06K9/6277G06K9/66G06K2009/3291
    • PROBLEM TO BE SOLVED: To provide an image processing device, an image processing method and a program capable of updating a template according to movement of an object.SOLUTION: An image processing device includes a target area specification part, a calculation part, a separation degree detection part, and a target area determination part. The target area specification part specifies a target area. The calculation part calculates, when an image is divided into multiple sub-areas each being smaller than the target area, the center of gravity obtained by weighting a central coordinate of each of the multiple sub-areas, with the likelihood of each of the multiple sub-areas, based on a histogram of feature values of the respective sub-areas inside and outside the target area, and a covariance matrix according to the center of gravity. The separation degree detection part detects the separation degree on the basis of the histogram, the likelihood, the center of gravity and the covariance matrix. The target area determination part determines an updated target area on the basis of the separation degree.
    • 要解决的问题:提供一种图像处理装置,图像处理方法和能够根据对象的移动来更新模板的程序。解决方案:图像处理装置包括目标区域指定部分,计算部分,分离 度检测部和目标区确定部。 目标区域规范部分指定目标区域。 计算部分计算当图像被划分成比目标区域小的多个子区域时,通过对多个子区域中的每个子区域的中心坐标进行加权而获得的重心, 基于目标区域内外的各个子区域的特征值的直方图以及根据重心的协方差矩阵的子区域。 分离度检测部分基于直方图,似然性,重心和协方差矩阵来检测分离度。 目标区域确定部分基于分离度来确定更新的目标区域。
    • 87. 发明专利
    • Image processing system, image processing method, and program
    • 图像处理系统,图像处理方法和程序
    • JP2013041398A
    • 2013-02-28
    • JP2011177525
    • 2011-08-15
    • Toshiba Corp株式会社東芝
    • KAWADA SATOSHIYAMAMOTO TAKUMATAGUCHI YASUNORIMATSUMOTO NOBUYUKI
    • G06T5/00G06T5/20H04N1/409
    • G06T5/002G06K9/66G06T5/20H04N1/409
    • PROBLEM TO BE SOLVED: To remove noise from an image while maintaining sharpness of the image.SOLUTION: An image processing system comprises a correlation calculation unit, a selection unit, a distance calculation unit, a feature amount calculation unit, a weight calculation unit, and a pixel value calculation unit. The correlation calculation unit calculates correlations between a first area including a first pixel in an image and a plurality of first base vectors. The selection unit selects one or more second base vectors from the first base vectors according to the magnitude of the correlations. The distance calculation unit calculates distances between the first area and a plurality of second areas each including a plurality of second pixels in a subspace formed by the second base vectors. The feature amount calculation unit calculates feature amounts by using the correlations corresponding to third base vectors of the first base vectors. The weight calculation unit calculates weight to be assigned to the second pixels according to the distances and the feature amounts. The pixel value calculation unit calculates output pixel values by calculating a weighted average of the second pixels according to the weight.
    • 要解决的问题:为了在保持图像清晰度的同时从图像中除去噪声。 解决方案:图像处理系统包括相关计算单元,选择单元,距离计算单元,特征量计算单元,权重计算单元和像素值计算单元。 相关计算单元计算包括图像中的第一像素的第一区域和多个第一基本矢量之间的相关性。 选择单元根据相关性的大小从第一基本向量中选择一个或多个第二基本向量。 距离计算单元计算第一区域和多个第二区域之间的距离,每个第二区域包括由第二基本矢量形成的子空间中的多个第二像素。 特征量计算单元通过使用与第一基矢量的第三基矢量相对应的相关来计算特征量。 权重计算单元根据距离和特征量计算要分配给第二像素的权重。 像素值计算单元通过根据权重计算第二像素的加权平均来计算输出像素值。 版权所有(C)2013,JPO&INPIT
    • 88. 发明专利
    • Object recognition device, control method of object recognition device and program
    • 对象识别装置,对象识别装置和程序的控制方法
    • JP2012238119A
    • 2012-12-06
    • JP2011105647
    • 2011-05-10
    • Canon Incキヤノン株式会社
    • YANO KOTARO
    • G06T7/00
    • G06K9/6262G06K9/66G06T7/248G06T2207/30201
    • PROBLEM TO BE SOLVED: To perform an object detection with high accuracy using an appropriate recognition model in processing of tracking an object in an image even when the tracking is stopped and resumed in a scene with a different background.SOLUTION: An object recognition device comprises: an acquisition part for sequentially acquiring images; an extraction part for extracting a partial area from an image and extracting a feature amount from the partial area; a recognition part for recognizing whether or not the partial area is a target object on the basis of the extracted feature amount, a first recognition model including a feature amount of a positive instance indicating the target object and a feature amount of a negative instance indicating a background of the target object, or a second recognition model including the feature amount of the positive instance; an update part for updating the first recognition model by adding the extracted feature amount thereto according to the recognition result; and an output part for outputting an object area recognized as the target object. The recognition part recognizes the partial area on the basis of the first recognition model when an object area of a previous image acquired by the acquisition part has been outputted, and recognizes the partial area on the basis of the second recognition model when the object area of the previous image has not been outputted.
    • 要解决的问题:即使在跟踪停止并在具有不同背景的场景中恢复的情况下,也可以使用适当的识别模型在跟踪图像中的对象的处理中以高精度执行对象检测。 解决方案:一种对象识别装置,包括:用于顺序地获取图像的获取部分; 提取部分,用于从图像中提取部分区域并从所述部分区域提取特征量; 基于所提取的特征量来识别部分区域是目标对象的识别部分,包括指示目标对象的正实例的特征量的第一识别模型和表示目标对象的否定实例的特征量 目标对象的背景,或包括正面实例的特征量的第二识别模型; 更新部分,用于根据识别结果将提取的特征量相加,来更新第一识别模型; 以及用于输出被识别为目标对象的对象区域的输出部。 当已经输出了由获取部分获取的先前图像的对象区域时,识别部分基于第一识别模型识别部分区域,并且当对象区域被识别时,基于第二识别模型识别部分区域 以前的图像尚未输出。 版权所有(C)2013,JPO&INPIT
    • 90. 发明专利
    • Reference data recognition/learning method and pattern recognition system
    • 参考数据识别/学习方法和模式识别系统
    • JP2005190429A
    • 2005-07-14
    • JP2003434596
    • 2003-12-26
    • Hiroshima Univ国立大学法人広島大学
    • MATTHAEUS HANSJUERGENKOIDE TETSUSHIMIZOGAMI MASAHIRO
    • G06N3/08G06F17/30G06K9/00G06K9/66G06T7/00
    • G06K9/66G06K9/00986Y10S707/99937
    • PROBLEM TO BE SOLVED: To provide a reference data recognition/learning method which automatically learns data of high appearance frequencies without requiring teacher data. SOLUTION: In processing (a), reference data being at the shortest distance to input data is detected as a winner. In processing (b), it is discriminated whether the distance between input data and the winner is shorter than a threshold or not. In processing (c), reference data detected as the winner is regarded as coinciding with input data and the rank of reference data is raised in the case that the distance between input data and the winner is shorter than the threshold. In processing (d), reference data detected as the winner is regarded as different from input data and input data is written in an associative memory as new reference data in the case that the distance between input data and the winner is equal to or longer than the threshold. In this case, reference data in higher ranks are stored for a longer period, and reference data in lower ranks are stored for a shorter period, and a rise width of the rank of reference data detected as the winner is properly set, and reference data is registered in a short-period storage or a long-period storage in accordance with a coincidence frequency. COPYRIGHT: (C)2005,JPO&NCIPI
    • 要解决的问题:提供一种参考数据识别/学习方法,其自动学习高出现频率的数据,而不需要教师数据。 解决方案:在处理(a)中,以与输入数据最短的距离的参考数据被检测为获胜者。 在处理(b)中,判别输入数据和胜者之间的距离是否短于阈值。 在处理(c)中,作为胜者检测到的参考数据被认为与输入数据一致,并且在输入数据和获胜者之间的距离小于阈值的情况下引起参考数据的等级。 在处理(d)中,作为获胜者检测到的参考数据被认为与输入数据不同,并且在输入数据和获胜者之间的距离等于或大于的情况下,输入数据被写入关联存储器中作为新的参考数据 门槛。 在这种情况下,较高级别的参考数据被存储较长时间段,较低级别的参考数据被存储较短的时间段,并且被正确地设置作为获胜者检测到的参考数据的等级的上升幅度,以及参考数据 根据符合频率被记录在短期存储或长周期存储中。 版权所有(C)2005,JPO&NCIPI