公开(公告)号：WO2016208037A1

公开(公告)日：2016-12-29

申请号：PCT/JP2015/068371

申请日：2015-06-25

Applicant: 富士通株式会社

Inventor： 長門　毅 ,  肥塚　哲男

IPC: G06N3/12 ,  G06F9/44 ,  G06T5/20 ,  G06T7/00

CPC classification number: G06N3/126 ,  G06F8/36 ,  G06F9/44 ,  G06K9/00993 ,  G06K9/6202 ,  G06K9/6229 ,  G06K9/6256 ,  G06N3/12 ,  G06T5/20 ,  G06T7/00

Abstract: 　遺伝的プログラミングによって画像処理プログラムを生成する際の生存選択を適正化する。　演算部（１ｂ）は、それぞれ複数の部分プログラムが組み合わされた複数の画像処理プログラム（２１，２２，２３，・・・）の中から画像処理プログラム（２１）を選択し、画像処理プログラム（２１）に含まれる部分プログラムの一部を変更することで画像処理プログラム（２１ａ）を生成し、画像処理プログラム（２１ａ）を用いて入力画像（１１）に対する画像処理を実行し、画像処理の途中で出力される中間出力画像（３１，３２）と第１目標画像（１２）との比較に基づいて、画像処理プログラム（２１ａ）を次世代に残すかを判定し、次世代に残すと判定された場合、画像処理プログラム（２１，２２，２３，・・・）の１つを画像処理プログラム（２１ａ）に入れ替える。

Abstract translation: 本发明的目的是通过遗传编程来生成图像处理程序时优化存活选择。 计算单元（1b）：从其中组合有多个部分程序的多个图像处理程序（21,22,23，...）中选择图像处理程序（21） 通过改变包括在图像处理程序（21）中的部分节目的一部分来生成图像处理程序（21a）。 使用图像处理程序（21a）对所输入的图像（11）执行图像处理; 基于在图像处理期间输出的第一目标图像（12）与中间输出图像（31,32）的比较，确定是否离开用于下一代的图像处理程序（21a）; 并且如果确定图像处理程序（21a）将被遗留下一代，则用图像处理程序（21a）替换图像处理程序（21,22,23，...）中的一个。

公开(公告)号：WO2015187401A1

公开(公告)日：2015-12-10

申请号：PCT/US2015/032374

申请日：2015-05-26

Applicant: ROTHMAN, Neil ,  JACQUIN, Arnaud ,  GHOSH DASTIDAR, Samanwoy ,  RADMAN, Thomas ,  MILLER, Jonathan ,  KOPPES, William ,  PRICHEP, Leslie

Inventor： ROTHMAN, Neil ,  JACQUIN, Arnaud ,  GHOSH DASTIDAR, Samanwoy ,  RADMAN, Thomas ,  MILLER, Jonathan ,  KOPPES, William ,  PRICHEP, Leslie

IPC: A61B5/00 ,  A61B5/16 ,  G06F19/00 ,  A61B5/0476 ,  G06K9/62

CPC classification number: A61B5/4064 ,  A61B5/0476 ,  A61B5/16 ,  A61B5/4082 ,  A61B5/4088 ,  A61B5/7267 ,  A61B2505/01 ,  A61B2505/09 ,  A61B2560/0487 ,  G06F19/00 ,  G06K9/6229 ,  G06K9/6267 ,  G16H50/20 ,  G16H50/30

Abstract: A method of building classifiers for neurological assessment is described. The method comprises the steps of extracting quantitative features from a plurality of clinical features, and selecting a subset of features from the extracted pool of features to construct binary classifiers. A device for performing point-of-care neurological assessment using clinical features is also described.

Abstract translation: 描述了建立神经学评估分类器的方法。 该方法包括以下步骤：从多个临床特征提取定量特征，以及从提取的特征库中选择特征子集以构建二进制分类器。 还描述了使用临床特征进行护理神经评估的装置。

公开(公告)号：WO2012098388A1

公开(公告)日：2012-07-26

申请号：PCT/GB2012/050093

申请日：2012-01-18

Applicant: THE UNIVERSITY OF YORK ,  SMITH, Stephen

Inventor： SMITH, Stephen

IPC: G06K9/00 ,  G06K9/62

CPC classification number: A61B5/1101 ,  A61B5/1125 ,  G06K9/00536 ,  G06K9/6229

Abstract: The invention provides signal processing algorithms and apparatus for detecting bradykinesia, tremor, or other symptoms of neurological dysfunction in subjects, using three-dimensional sensors to tract finger and hand position. The invention provides Cartesian Genetic Programming networks and particular function blocks for such networks to enable identification of subjects exhibiting such symptoms.

Abstract translation: 本发明提供用于检测受试者的运动障碍，震颤或其他神经功能障碍症状的信号处理算法和装置，其使用三维传感器来管理手指和手部位置。 本发明提供了笛卡尔基因编程网络和用于这种网络的特定功能块，以使得能够识别表现出这种症状的受试者。

公开(公告)号：WO2006004797A2

公开(公告)日：2006-01-12

申请号：PCT/US2005/023018

申请日：2005-06-27

Applicant: THE TRUSTEES OF COLUMBIA UNIVERSITY IN THE CITY OFNEW YORK ,  LIU, Yan ,  KENDER, John, R.

Inventor： LIU, Yan ,  KENDER, John, R.

IPC: G06K9/62

CPC classification number: G06K9/6228 ,  G06K9/6229 ,  Y10S707/99937

Abstract: Methods and systems for feature selection are described. In particular, methods and systems for feature selection for data classification, retrieval, and segmentation are described. Certain embodiments of the invention are directed to methods and systems for complement sort-merge tree (CSMT), fast-converging sort-merge tree (FSMT), and multi-level (ML) feature selection. Accurate and fast results may be obtained by the feature selection methods and systems described herein.

Abstract translation: 描述用于特征选择的方法和系统。 特别地，描述了用于数据分类，检索和分割的特征选择的方法和系统。 本发明的某些实施例涉及用于补码分类合并树（CSMT），快速聚合分类合并树（FSMT）和多级（ML）特征选择的方法和系统。 通过本文描述的特征选择方法和系统可以获得准确和快速的结果。

公开(公告)号：WO2016118332A1

公开(公告)日：2016-07-28

申请号：PCT/US2016/012522

申请日：2016-01-07

Applicant: EBAY INC.

Inventor： KIAPOUR, M. Hadi ,  DI, Wei ,  JAGADEESH, Vignesh ,  PIRAMUTHU, Robinson

IPC: B41L47/58 ,  G06K9/60 ,  G06T3/60 ,  G06T7/00 ,  H04N1/387

CPC classification number: G06K9/6276 ,  G06K9/00979 ,  G06K9/3216 ,  G06K9/3241 ,  G06K9/4628 ,  G06K9/6229 ,  G06K9/6256 ,  G06K9/6857 ,  G06Q30/00

Abstract: An image is passed through an image identifier to identify a coarse category for the image and a bounding box for a categorized object. A mask is used to identify the portion of the image that represents the object. Given the foreground mask, the convex hull of the mask is located and an aligned rectangle of minimum area that encloses the hull is fitted. The aligned bounding box is rotated and scaled, so that the foreground object is roughly moved to a standard orientation and size (referred to as calibrated). The calibrated image is used as an input to a fine-grained categorization module, which determines the fine category within the coarse category for the input image.

Abstract translation: 图像通过图像标识符，以识别图像的粗略类别和分类对象的边界框。 使用掩码来识别代表对象的图像部分。 给定前景掩模，掩模的凸包被定位，并且装配包围船体的最小面积的排列的矩形。 对齐的边界框被旋转和缩放，使得前景对象粗略地移动到标准方向和大小（称为校准）。 校准图像用作细粒度分类模块的输入，该模块确定输入图像的粗略类别内的精细类别。

公开(公告)号：WO2014195782A2

公开(公告)日：2014-12-11

申请号：PCT/IB2014/000939

申请日：2014-06-03

Applicant: TATA CONSULTANCY SERVICES LIMITED

Inventor： CHAKRAVARTY, Kingshuk ,  DAS, Diptesh ,  SINHA, Aniruddha ,  KONAR, Amit

IPC: G06F17/40

CPC classification number: G06K9/00523 ,  A61B5/0476 ,  A61B5/7264 ,  G06K9/6229 ,  G06K9/6231

Abstract: The subject matter discloses systems and methods for selection of an optimum feature subset. According to the present subject matter, the system (102) implements the described method, where the method includes obtaining a plurality of features extracted from data sets associated with objects representing multiple classes, computing an intra-class variation factor and an inter-class variation factor for multiple feature subsets, from amongst the plurality of features, and identifying an optimum feature subset, from amongst the multiple feature subsets, based on minimization of the intra-class variation factor and maximization of the inter-class variation factor using differential evolution.

Abstract translation: 主题公开了用于选择最佳特征子集的系统和方法。 根据本主题，系统（102）实现所述方法，其中该方法包括获得从与表示多个类的对象相关联的数据集中提取的多个特征，计算类内变化因子和类间变化 基于来自多个特征的多个特征子集的因子，以及从多个特征子集中，基于类内变异因子的最小化和使用差分进化的类间变异因子的最大化来识别最佳特征子集。

公开(公告)号：WO2007099495A1

公开(公告)日：2007-09-07

申请号：PCT/IB2007/050621

申请日：2007-02-27

Applicant: KONINKLIJKE PHILIPS ELECTRONICS N.V. ,  ZHAO, Luyin

Inventor： ZHAO, Luyin

IPC: G06K9/62

CPC classification number: G06K9/6229

Abstract: The invention relates to a method (100) of and a system (200) for identifying a set of image characteristics for assessing similarity of images from a pool of image characteristics on the basis of a set of training images. The obtained set of image characteristics is especially useful for identifying images depicting similar objects. Advantageously, the identified set of image characteristics is human-oriented in the sense that it is based on human perception of image similarity thanks to the use of human rating as a reference for the machine rating of similarity of images. The invention further relates to a method of and a system for identifying a reference image from a database of images on the basis of similarity of the reference image to a given image using the set of image characteristics.

Abstract translation: 本发明涉及一种用于基于一组训练图像来识别用于评估来自图像特征池的图像的相似性的一组图像特征的系统（200）和系统（200）。 所获得的图像特征集对于识别描绘相似对象的图像特别有用。 有利的是，鉴于一组图像特征是以人为本的意义上的，因为它基于人类对图像相似性的感知，这是由于使用人类等级作为图像相似度的机器评级的参考。 本发明还涉及一种用于使用该组图像特性基于参考图像与给定图像的相似性来从图像数据库识别参考图像的方法和系统。

公开(公告)号：WO2006004797A3

公开(公告)日：2007-03-01

申请号：PCT/US2005023018

申请日：2005-06-27

Applicant: UNIV COLUMBIA ,  LIU YAN ,  KENDER JOHN R

Inventor： LIU YAN ,  KENDER JOHN R

IPC: G06K9/62

CPC classification number: G06K9/6228 ,  G06K9/6229 ,  Y10S707/99937

Abstract: Methods and systems for feature selection are described. In particular, methods and systems for feature selection for data classification, retrieval, and segmentation are described. Certain embodiments of the invention are directed to methods and systems for complement sort-merge tree (CSMT), fast-converging sort-merge tree (FSMT), and multi-level (ML) feature selection. Accurate and fast results may be obtained by the feature selection methods and systems described herein.

Abstract translation: 描述用于特征选择的方法和系统。 特别地，描述了用于数据分类，检索和分割的特征选择的方法和系统。 本发明的某些实施例涉及用于补码分类合并树（CSMT），快速聚合分类合并树（FSMT）和多级（ML）特征选择的方法和系统。 通过本文描述的特征选择方法和系统可以获得准确和快速的结果。

公开(公告)号：WO1997044741A1

公开(公告)日：1997-11-27

申请号：PCT/US1997008951

申请日：1997-05-23

Applicant: THE UNITED STATES OF AMERICA, represented by THE SECRETARY, DEPARTMENT OF HEALTH AND HUMAN SERVICES, THE NATIONAL INSTITUTES OF HEALTH ,  HUNTER, Lawrence, E.

Inventor： THE UNITED STATES OF AMERICA, represented by THE SECRETARY, DEPARTMENT OF HEALTH AND HUMAN SERVICES, THE NATIONAL INSTITUTES OF HEALTH

IPC: G06F15/18

CPC classification number: G06K9/6229 ,  G06N3/004 ,  G06N5/043

Abstract: System and method for improving the performance of learning agents such as neural networks, genetic algorithms and decision trees that derive prediction methods from a training set of data. In part of the method, a population of learning agents of different classes is trained on the data set, each agent producing in response a prediction method based on the agent's input representation. Feature combinations are extracted from the prediction methods produced by the learning agents. The input representations of the learning agents are then modified by including therein a feature combination extracted from another learning agent. In another part of the method, the parameter values of the learning agents are changed to improve the accuracy of the prediction method. A fitness measure is determined for each learning agent based on the prediction method the agent produces. Parameter values of a learning agent are then selected based on the agent's fitness measure. Variation is introduced into the selected parameter values, and another learning agent of the same class is defined using the varied parameter values. The learning agents are then again trained on the data set to cause a learning agent to produce a prediction method based on the derived feature combinations and varied parameter values.

Abstract translation: 用于提高诸如神经网络，遗传算法和决策树等学习代理的性能的系统和方法，其从训练数据集中导出预测方法。 在该方法的一部分中，对数据集合训练不同类的学习代理人群，每个代理响应于基于代理的输入表示的预测方法产生。 从学习者产生的预测方法中提取特征组合。 然后通过在其中包括从另一个学习代理提取的特征组合来修改学习代理的输入表示。 在该方法的另一部分中，改变学习者的参数值以提高预测方法的准确性。 基于代理产生的预测方法，针对每个学习代理确定适应性度量。 然后基于代理的适应度度量来选择学习代理的参数值。 将变化引入到所选择的参数值中，并且使用变化的参数值来定义同一类的另一个学习代理。 然后再次对数据集训练学习代理，以使学习代理基于导出的特征组合和变化的参数值产生预测方法。

公开(公告)号：WO2019122271A1

公开(公告)日：2019-06-27

申请号：PCT/EP2018/086442

申请日：2018-12-20

Applicant: ALPVISION S.A.

Inventor： JORDAN, Frederic ,  RUDAZ, Nicolas ,  DELACRETAZ, Yves ,  KUTTER, Martin

IPC: G06K9/62

CPC classification number: G06K9/6227 ,  G06K9/62 ,  G06K9/6228 ,  G06K9/6229

Abstract: A machine learning system may automatically produce classifier algorithms and configuration parameters by selecting them into a set of predetermined unitary algorithms and associated parametrization values. Multiple digital representations of input object items may be produced by varying the position and orientation of the object to be classified and/or of the sensor to capture a digital representation of the object, and/or by varying a physical environment parameter which changes the digital representation capture of the object by the sensor. A robot arm or a conveyor may vary the object and/or the sensor positions and orientations. The machine learning system may employ genetic programming to facilitate the production of classifiers suitable for the classification of multiple digital representations of input object items. The machine learning system may automatically generate reference template signals as configuration parameters for the unitary algorithms to facilitate the production of classifiers suitable for the classification of multiple digital representations of input object items.