公开(公告)号：US20190108423A1

公开(公告)日：2019-04-11

申请号：US15726415

申请日：2017-10-06

申请人： Mitsubishi Electric Research Laboratories, Inc.

发明人： Michael Jones

IPC分类号： G06K9/62 ,  G06K9/00

CPC分类号： G06K9/6286 ,  G06K9/00288 ,  G06K9/6215 ,  G06K9/6269

摘要： An image processing system includes a memory to store data indicative of benchmark feature vectors of benchmark images, an input interface to accept data indicative of a first feature vector of a first image and a second feature vector of a second image, and an output interface to render a similarity value between the first and the second images. The system includes a processor to determine the similarity using a first hyperplane separating the benchmark feature vectors from the first feature vector and a second hyperplane separating the benchmark feature vectors from the second feature vector. The processor determines a first normal vector to the first hyperplane as the difference between the first feature vector and the mean of the benchmark feature vectors. The processor determines an offset for the first hyperplane as the average of the maximum inner product of the benchmark feature vectors with the first normal vector and the inner product of the first feature vector with the first normal vector. The processor determines the similarity value as a function of a sum of a signed distance of the second feature vector to the first hyperplane and a signed distance of the first feature vector to the second hyperplane.

公开(公告)号：US09602783B2

公开(公告)日：2017-03-21

申请号：US14730154

申请日：2015-06-03

申请人： Panasonic Intellectual Property Management Co., Ltd.

发明人： Yoshihiro Sugishita ,  Ryohei Wakai ,  Akihiro Miyazaki

IPC分类号： G06K9/00 ,  H04N7/18 ,  G06K9/62 ,  H04N5/217

CPC分类号： H04N7/181 ,  G06K9/00771 ,  G06K9/00993 ,  G06K9/6257 ,  G06K9/6286 ,  H04N7/188

摘要： A first image taken by a first camera device in the plurality of camera devices and first imaging environment information indicating a first imaging environment of the first camera device at a time of taking the first image is acquired. By using a parameter table that manages imaging environment information indicating an imaging environment at a time of taking an image previously by a camera device and a recognition control parameter indicating a detector corresponding to an imaging environment, a first recognition control parameter indicating a first detector corresponding to third imaging environment that is identical or similar to the first imaging environment indicated by the first imaging environment information acquired from the first camera device is selected from the recognition control parameters. The first image acquired from the first camera device is recognized by using the first detector indicated by the selected first recognition control parameter.

公开(公告)号：US20170076466A1

公开(公告)日：2017-03-16

申请号：US15363032

申请日：2016-11-29

申请人： IRIDA LABS S.A.

发明人： Ilias Theodorakopoulos ,  Nikos Fragoulis

IPC分类号： G06T7/20 ,  G06K9/00 ,  G06K9/62

CPC分类号： G06K9/627 ,  G06K9/00335 ,  G06K9/6272 ,  G06K9/6277 ,  G06K9/6286 ,  G06T7/246 ,  G06T2207/10016

摘要： In the event that a moving body (e.g. a person, a car, etc.) is outfitted with a video camera or with a camera-equipped device (e.g. a tablet or a mobile phone), the system described in one aspect is able to understand the motion of the moving by analyzing the video frame sequence captured by the camera. This means that the system can categorize the motion of the body-carrying camera to one of several types (e.g., is this a person walking? is this a person running? etc.), understand the nature of the moving body holding the camera-equipped device (e.g. Is this a car?, Is this a person? etc.) and even to identify the moving body (which car?, which person? etc.).

摘要翻译： 如果移动体（例如，人，汽车等）配备有摄像机或配备有相机的设备（例如，平板电脑或移动电话），则一方面描述的系统能够 通过分析摄像机拍摄的视频帧序列来了解移动的动作。 这意味着系统可以将身体携带摄像机的运动分类为几种类型之一（例如，这个人是走路的，这是一个人在跑吗？等等），了解保持相机的移动体的性质 - 装备的装置（例如这是一辆车，这是一个人吗？等等），甚至识别移动体（哪个车，哪个人？等）。

公开(公告)号：US20160321521A1

公开(公告)日：2016-11-03

申请号：US15139603

申请日：2016-04-27

申请人： CANON KABUSHIKI KAISHA

发明人： Tsewei Chen

IPC分类号： G06K9/62 ,  G06K9/00

CPC分类号： G06K9/6282 ,  G06K9/00228 ,  G06K9/6257 ,  G06K9/6286

摘要： In an information processing apparatus that includes sequences of weak classifiers which are logically cascade-connected in each sequence and the sequences respectively correspond to categories of an object and in which the weak classifiers are grouped into at least a first group and a second group in the order of connection, classification processing by weak classifiers belonging to the first group of respective categories is performed by pipeline processing. Based on the processing results of the weak classifiers belonging to the first group of the respective categories, categories in which classification processing by weak classifiers belonging to the second group is to be performed are decided out of the categories. The classification processing by the weak classifiers respectively corresponding to the decided categories and belonging to the second group is performed by pipeline processing.

摘要翻译： 在包括在每个序列中逻辑级联连接的弱分类器的序列的信息处理设备中，并且所述序列分别对应于对象的类别，并且其中所述弱分类器被分组为至少第一组和 通过流水线处理执行属于第一组各自类别的弱分类器的连接顺序，分类处理。 根据属于第一组各类别的弱分类器的处理结果，根据属于第二组的弱分类器进行分类处理的类别进行判定。 分类对应于属于第二组的决定类别的弱分类器的分类处理通过流水线处理来执行。

公开(公告)号：US20110137841A1

公开(公告)日：2011-06-09

申请号：US13019683

申请日：2011-02-02

申请人： Kohtarou YUTA

发明人： Kohtarou YUTA

IPC分类号： G06F15/18

CPC分类号： G06K9/6286 ,  G06F19/24 ,  G06F19/704 ,  G06F19/707 ,  G06K9/6234 ,  G06K9/6256

摘要： To predict the class of an unknown sample, a) a discriminant function for assigning each training sample to class 1 or class 2 is obtained, b) the discriminant score of each training sample and an unknown sample are calculated using the function, c) it is determined whether the score of the unknown sample is either not smaller than the largest score or not larger than the smallest score taken among all of the training samples, d) if the determination in c) is affirmative, the class of the unknown sample is determined based on the score of the unknown sample, e) if the determination in c) is negative, then the training samples having the largest score and the smallest score are removed to form a new training sample set from remaining training samples, and f) a) to e) are repeated.

摘要翻译： 为了预测未知样本的类别，a）获得将每个训练样本分配给1类或2类的判别函数，b）使用该函数计算每个训练样本和未知样本的判别分数，c） 确定未知样本的得分是否不小于最大得分或不大于所有训练样本中取得的最小得分，d）如果c）中的确定是肯定的，则未知样本的类别是 根据未知样本的得分确定，e）如果c）中的确定为否定，则删除具有最大得分和最小得分的训练样本，以形成剩余训练样本的新训练样本集，f） a）至e）重复。

公开(公告)号：US07848558B2

公开(公告)日：2010-12-07

申请号：US10777041

申请日：2004-02-13

申请人： Maryellen L. Giger ,  Hui Li

发明人： Maryellen L. Giger ,  Hui Li

IPC分类号： G06K9/00

CPC分类号： G06K9/6286 ,  G06T7/0012 ,  G06T7/48 ,  G06T2207/20104 ,  G06T2207/30068

摘要： A computerized method, system and computer program for the computerized fractal-based analysis of a structure as presented in a pattern on a medical image. Image data is generated from the medical image and a region of interest is selected. The image data is digitized and analyzed to reveal fractal-based computer-generated features of a texture of the image data. Then a qualifier is applied to the computer-generated features to obtain fractal characteristics of the image data. A multi-fractal nature is observed for the texture of the region of interest. A marker for assessing a risk of a disease is yielded based on the multi-fractal nature of the texture.

摘要翻译： 用于以医学图像上的图案呈现的结构的基于计算机化的基于分形的分析的计算机化方法，系统和计算机程序。 从医学图像生成图像数据，并且选择感兴趣的区域。 数字化和分析图像数据以显示图像数据纹理的基于分形的计算机生成的特征。 然后，将限定符应用于计算机生成的特征以获得图像数据的分形特征。 观察到感兴趣区域的纹理的多分形性质。 基于纹理的多分形性质产生用于评估疾病风险的标记。

公开(公告)号：US20100074538A1

公开(公告)日：2010-03-25

申请号：US12237817

申请日：2008-09-25

申请人： Pragyana K. Mishra ,  Eyal Ofek ,  Gur Kimchi

发明人： Pragyana K. Mishra ,  Eyal Ofek ,  Gur Kimchi

IPC分类号： G06K9/62

CPC分类号： G06K9/6286 ,  G06T7/0002 ,  G06T17/05 ,  G06T2207/10032

摘要： Technologies are described herein for validating and correcting map data using oblique images or aerial photographs taken at oblique angles to the earth's surface. Pixels within oblique images can be analyzed to detect, validate, and correct other sources of data used in generating maps such as vector data, elevation maps, projection parameters, and three-dimensional model data. Visibility and occlusion information in oblique views may be analyzed to reduce errors in either occluding or occluded entities. Occlusion of road segments due to foliage, z-ordering of freeways, tunnels, bridges, buildings, and other geospatial entities may be determined, validated, and corrected. A learning algorithm can be trained with image-based descriptors that encode visible data consistencies. After training, the algorithm can classify errors and inconsistencies using combinations of different descriptors such as color, texture, image-gradients, and filter responses.

摘要翻译： 这里描述了用于使用与地球表面倾斜的角度拍摄的倾斜图像或航空照片来验证和校正地图数据的技术。 可以分析倾斜图像中的像素以检测，验证和校正生成地图中使用的其他数据源，如矢量数据，高程图，投影参数和三维模型数据。 可以分析斜视图中的可视性和遮挡信息，以减少闭塞或闭塞实体中的误差。 可以确定，验证和纠正由于树叶，高速公路，隧道，桥梁，建筑物和其他地理空间实体造成的路段堵塞。 可以使用编码可视数据一致性的基于图像的描述符来训练学习算法。 训练后，该算法可以使用不同描述符（如颜色，纹理，图像梯度和过滤器响应）的组合来分类错误和不一致。

公开(公告)号：US20040258310A1

公开(公告)日：2004-12-23

申请号：US10777041

申请日：2004-02-13

申请人： The University of Chicago

发明人： Maryellen L. Giger ,  Hui Li

IPC分类号： G06K009/62 ,  G06K009/00

CPC分类号： G06K9/6286 ,  G06T7/0012 ,  G06T7/48 ,  G06T2207/20104 ,  G06T2207/30068

摘要： A computerized method, system and computer program for the computerized fractal-based analysis of a structure as presented in a pattern on a medical image. Image data is generated from the medical image and a region of interest is selected. The image data is digitized and analyzed to reveal fractal-based computer-generated features of a texture of the image data. Then a qualifier is applied to the computer-generated features to obtain fractal characteristics of the image data. A multi-fractal nature is observed for the texture of the region of interest. A marker for assessing a risk of a disease is yielded based on the multi-fractal nature of the texture.

摘要翻译： 用于以医学图像上的图案呈现的结构的基于计算机化的基于分形的分析的计算机化方法，系统和计算机程序。 从医学图像生成图像数据，并且选择感兴趣的区域。 数字化和分析图像数据以显示图像数据纹理的基于分形的计算机生成的特征。 然后，将限定符应用于计算机生成的特征以获得图像数据的分形特征。 观察到感兴趣区域的纹理的多分形性质。 基于纹理的多分形性质产生用于评估疾病风险的标记。

公开(公告)号：US06571225B1

公开(公告)日：2003-05-27

申请号：US09502578

申请日：2000-02-11

申请人： Frank J. Oles ,  Tong Zhang

发明人： Frank J. Oles ,  Tong Zhang

IPC分类号： G06F1518

CPC分类号： G06N99/005 ,  G06K9/6286 ,  Y10S707/99937

摘要： A method to automatically categorize messages or documents containing text. The method of solution fits in the general framework of supervised learning, in which a rule or rules for categorizing data is automatically constructed by a computer on the basis of training data that has been labeled beforehand. More specifically, the method involves the construction of a linear separator: training data is used to construct for each category a weight vector w and a threshold t, and the decision of whether a hitherto unseen document d is in the category will depend on the outcome of the test wTx≧t, where x is a vector derived from the document d. The method also uses a set L of features selected from the training data in order to construct the numerical vector representation x of a document. The preferred method uses an algorithm based on Gauss-Seidel iteration to determine the weight factor w that is determined by a regularized convex optimization problem derived from the principle of minimizing modified training error.

摘要翻译： 一种自动分类包含文本的消息或文档的方法。 解决方法适合于监督学习的一般框架，其中用于分类数据的规则或规则由计算机根据预先标注的训练数据自动构建。 更具体地说，该方法涉及线性分离器的构造：训练数据用于为每个类别构建权重向量w和阈值t，并且决定迄今为止看不到的文档d是否在类别中将取决于结果 的测试wTx> = t，其中x是从文档导出的向量d。 该方法还使用从训练数据中选择的一组特征集合来构造文档的数字矢量表示x。 优选方法使用基于Gauss-Seidel迭代的算法来确定由从最小化修改的训练误差的原理导出的正则化凸优化问题确定的权重因子w。

公开(公告)号：US3725928A

公开(公告)日：1973-04-03

申请号：US3725928D

申请日：1971-03-31

申请人： BENDIX CORP

发明人： KLIMCHAK E ,  ROSENBAUM E

IPC分类号： G01S7/41 ,  G01S9/02

CPC分类号： G06K9/6286 ,  G01S7/412 ,  G06K9/00986

摘要： A hyperplane radar signature recognizer consists of a delay line through which a radar signature which is to be classified is propagated. The delay line includes a plurality of time spaced taps at which time spaced points on the radar signature are sampled. Each delay line tap is sampled through an associated electrically controlled weight with the weighted samples being summed to provide a measure of signal classification. The electrically controlled weights are controlled by propagating through the delay line prior to the receipt of the radar signature an amplitude standardized pulse. With all weights disabled so as to present an open circuit to the aforementioned summing network the weights are enabled one at a time in turn, and the output from the summing network compared against a desired weight value. The electrically controlled weight is then adjusted until its value is equal to the desired value. After all the electrically controlled weights have been set to their desired values the signature recognizer is ready to receive the radar signature. The timing logic circuit required to program the adjustment of the electrically controlled weights is also disclosed.

摘要翻译： 超平面雷达签名识别器由延迟线组成，通过该延迟线传播要分类的雷达签名。 延迟线包括多个时间间隔的抽头，在雷达签名上的时间间隔点被采样。 每个延迟线抽头通过相关的电控重量进行采样，加权采样相加以提供信号分类的度量。 电子控制权重通过在接收到雷达签名之前通过延迟线传播振幅标准化脉冲来控制。 在所有权重被禁用以便向上述求和网络呈现开路的情况下，权重依次被一个一个地启用，并且来自求和网络的输出与期望的权重值进行比较。 然后调整电控重量，直到其值等于所需值。 在所有的电控重量已被设置为其期望值之后，签名识别器准备好接收雷达签名。 还公开了对电控重量的调整进行编程所需的定时逻辑电路。