公开(公告)号：US20230162001A1

公开(公告)日：2023-05-25

申请号：US18058433

申请日：2022-11-23

Applicant: SEIKO EPSON CORPORATION

Inventor： Shin NISHIMURA ,  Hikaru KURASAWA

IPC: G06N3/04

CPC classification number: G06N3/04

Abstract: A classification device executes classification processing for data to be classified using a machine learning model including a vector neural network including a plurality of vector neuron layers. The machine learning model includes an input layer, an intermediate layer, and a first output layer and a second output layer that are branched from the intermediate layer, the first output layer is configured to use a first activation function, and the second output layer is configured to use a second activation function that is different from the first activation function.

公开(公告)号：US20160103018A1

公开(公告)日：2016-04-14

申请号：US14877827

申请日：2015-10-07

Applicant: SEIKO EPSON CORPORATION

Inventor： Hikaru KURASAWA ,  Yoshifumi ARAI

IPC: G01J3/02 ,  G01J3/28

CPC classification number: G01J3/0297 ,  G01J3/28 ,  G01N21/359 ,  G01N2201/1293 ,  G06F17/18 ,  G06K9/2018 ,  G06K9/6242

Abstract: A calibration curve generation method includes acquiring an independent component matrix including independent components of each sample the independent component matrix by performing first pre-processing that includes normalizing observation data, second pre-processing that includes whitening, and independent component analysis processing in order. Further, the same noise is added to the observation data related to a plurality of samples in the first pre-processing.

Abstract translation: 校准曲线生成方法包括通过执行包括归一化观察数据，包括白化的第二预处理和独立分量分析处理的第一预处理来获取包括每个样本的独立分量的独立分量矩阵的独立分量矩阵。 此外，在第一预处理中，与多个采样相关的观测数据相加噪声。

公开(公告)号：US20160097676A1

公开(公告)日：2016-04-07

申请号：US14858462

申请日：2015-09-18

Applicant: SEIKO EPSON CORPORATION

Inventor： Hikaru KURASAWA ,  Yoshifumi ARAI

IPC: G01J3/02 ,  G01J3/42

CPC classification number: G01J3/0297 ,  A61B5/0075 ,  A61B5/01 ,  A61B5/14532 ,  A61B5/1455 ,  A61B5/1495 ,  A61B2560/0223 ,  A61B2560/0247 ,  A61B2560/0252 ,  G01J3/42 ,  G01K13/002 ,  G01N21/359 ,  G01N2201/1293 ,  G06F17/18 ,  G06K9/2018 ,  G06K9/624 ,  G06K9/6242

Abstract: A target component calibration device includes a mixing coefficient calculating section that calculates mixing coefficients of target components regarding a test object based on observational data regarding the test object and calibration data, and a target component content calculating section that calculates the content of target components based on the mixing coefficients calculated by the mixing coefficient calculating section and a simple regression equation representing the relationship between the content and the mixing coefficients corresponding to the target components. The target component content calculating section adjusts at least one of two constants of the simple regression equation depending on a measurement condition when the observational data is obtained.

Abstract translation: 目标部件校准装置包括：混合系数计算部，其基于与测试对象和校准数据有关的观测数据，计算关于测试对象的目标分量的混合系数;以及目标分量内容计算部，其基于 由混合系数计算部分计算的混合系数和表示内容与对应于目标分量的混合系数之间的关系的简单回归方程式。 目标分量内容计算部根据获得观测数据时的测量条件来调整简单回归方程的两个常数中的至少一个。

公开(公告)号：US20140297197A1

公开(公告)日：2014-10-02

申请号：US14224872

申请日：2014-03-25

Applicant: Seiko Epson Corporation

Inventor： Hikaru KURASAWA ,  Yoshifumi ARAI

IPC: G01N33/00

CPC classification number: G01N33/0098 ,  G01N21/27 ,  G01N2201/129

Abstract: A calibration curve creation method is capable of performing accurate measurement from a piece of observation data. The calibration curve creation method includes (a) acquiring observation data regarding a plurality of samples of a subject, (b) acquiring the content of a target component in each sample, (c) estimating a plurality of independent components at the time of separation into a plurality of independent components of each sample and calculating a mixing coefficient corresponding to the target component for each sample, and (d) calculating the regression equation of the calibration curve. The process (c) includes a step of calculating an independent component matrix by executing first pre-processing including correcting the observation data, second pre-processing including whitening, and independent component analysis processing in this order. A process suitable for the observation data is selected from a plurality of processes, and is used as the first pre-processing and the second pre-processing.

Abstract translation: 校准曲线创建方法能够从一条观察数据执行准确的测量。 校准曲线生成方法包括：（a）获取关于被检体的多个样本的观察数据，（b）获取各样本中的目标成分的含量，（c）在分离时估计多个独立成分 每个样本的多个独立分量，并计算与每个样本的目标分量对应的混合系数，以及（d）计算校准曲线的回归方程。 方法（c）包括通过执行包括校正观察数据的第一预处理，包括白化的第二预处理和独立分量分析处理来依次计算独立分量矩阵的步骤。 从多个处理中选择适合于观察数据的处理，并且将其用作第一预处理和第二预处理。

公开(公告)号：US20230169307A1

公开(公告)日：2023-06-01

申请号：US18058862

申请日：2022-11-26

Applicant: SEIKO EPSON CORPORATION

Inventor： Tomomasa USUI ,  Ryoki WATANABE ,  Hikaru KURASAWA ,  Shin NISHIMURA

IPC: G06N3/04 ,  G06V10/774

CPC classification number: G06N3/04 ,  G06V10/774

Abstract: A method according to the present disclosure includes (a) generating N pieces of input data from one target object, (b) inputting the input data to a machine learning model and obtaining M classification output values, one determination class, and a feature spectrum, (c) obtaining a similarity degree between a known feature spectrum group and the feature spectrum for the input data, and obtaining a reliability degree with respect to the determination class as a function of the reliability degree, and (d) executing a vote for the determination class, based on the reliability degree with respect to the determination class, and determining a class determination result of the target object, based on a result of the vote.

公开(公告)号：US20220164658A1

公开(公告)日：2022-05-26

申请号：US17535256

申请日：2021-11-24

Applicant: SEIKO EPSON CORPORATION

Inventor： Kana KANAZAWA ,  Hikaru KURASAWA ,  Ryoki WATANABE

IPC: G06N3/08 ,  G06N3/04 ,  G06V10/82 ,  G06V10/74

Abstract: A method causes one or more processors to execute a method in which a machine learning model of a vector neural network type is used. The model is learned to reproduce correspondence between first images and a pre-label corresponding to each of the first images, and includes one or more neuron layers. First intermediate data output by the one or more neurons when the first images are input to the learned model is stored in one or more memories in correlation with the neurons. The method includes inputting a second image of an object to the machine learning model and acquiring second intermediate data based on at least one of a second vector and a second activation included in the one or more neurons, calculating a similarity degree between the first and second intermediate data, generating an evidence image corresponding to the similarity degree, and displaying the generated evidence image.

公开(公告)号：US20220164577A1

公开(公告)日：2022-05-26

申请号：US17535141

申请日：2021-11-24

Applicant: SEIKO EPSON CORPORATION

Inventor： Hikaru KURASAWA

IPC: G06K9/00 ,  G06N3/04 ,  G06N3/063 ,  G06K9/62 ,  G06K9/46

Abstract: An object detection method includes inputting an input image to a learned machine learning model and generating a similarity image from an output of at least one specific layer, and generating a discriminant image to which at least an unknown label is assigned, by comparing a similarity of each pixel in the similarity image to a predetermined threshold value.

公开(公告)号：US20220138526A1

公开(公告)日：2022-05-05

申请号：US17514350

申请日：2021-10-29

Applicant: SEIKO EPSON CORPORATION

Inventor： Ryoki WATANABE ,  Hikaru KURASAWA

IPC: G06N3/04

Abstract: A method of making a single processor or a plurality of processors perform classification processing of classification target data using a machine learning model includes the steps of (a) preparing N machine learning models in a memory assuming N as an integer no smaller than 2, and (b) performing the classification processing of the classification target data using the N machine learning models. Each of the N machine learning models is configured so as to classify input data into any of a plurality of classes with learning using training data, and is configured so as to have at least one class different from a class of another of the N machine learning models.

公开(公告)号：US20200242808A1

公开(公告)日：2020-07-30

申请号：US16774101

申请日：2020-01-28

Applicant: Seiko Epson Corporation

Inventor： Ryoki WATANABE ,  Hikaru KURASAWA ,  Naoki KUWATA ,  Masashi KANAI

IPC: G06T7/90 ,  G01J3/28 ,  G06Q30/00

Abstract: An identification method for identifying a target to be measured includes accepting, from an input section, an information representing a condition for acquiring spectral information specific to a target to be measured, capturing, by a spectrometry camera, an image of the target, acquiring the spectral information specific to the target based on the captured image, and identifying the target based on (i) the spectral information and (ii) a database, stored in a memory, containing a plurality of pieces of spectral information corresponding to a plurality of objects. Acquiring the spectral information includes preferentially acquiring the spectral information specific to the target in a specific wavelength region where the target is identifiable.

公开(公告)号：US20230186047A1

公开(公告)日：2023-06-15

申请号：US18063897

申请日：2022-12-09

Applicant: SEIKO EPSON CORPORATION

Inventor： Hikaru KURASAWA ,  Yuki URUSHIBATA ,  Ryoki WATANABE ,  Shin NISHIMURA ,  Eiichiro YAMAGUCHI

IPC: G06N3/04 ,  G06N3/08

CPC classification number: G06N3/04 ,  G06N3/08

Abstract: An evaluation method for a trained machine learning model includes the steps of (a) inputting evaluation data to the trained machine learning model to generate first explanatory information used for an evaluation of the machine learning model, (b) using a value indicated by each piece of information included in the first explanatory information to generate second explanatory information indicating an evaluation of the trained machine learning model, and (c) outputting the generated second explanatory information.