公开(公告)号：US11929174B2

公开(公告)日：2024-03-12

申请号：US16985237

申请日：2020-08-05

Applicant: FUJIFILM Corporation

Inventor： Yoshiro Kitamura

IPC: G06N3/08 ,  G06F18/21 ,  G06F18/2431 ,  G06N3/045 ,  G06V10/44 ,  G06V10/764 ,  G06V10/82 ,  G16H50/20 ,  G16H30/40

CPC classification number: G16H50/20 ,  G06F18/217 ,  G06F18/2431 ,  G06N3/045 ,  G06N3/08 ,  G06V10/454 ,  G06V10/764 ,  G06V10/82 ,  G06V2201/031 ,  G16H30/40

Abstract: A machine learning method and an apparatus, a program, a learned model, and a discrimination apparatus capable of controlling a calculation amount by learning a new task without changing output performance for an existing task in a learned network are provided. A machine learning method according to one aspect of the present disclosure includes a step of adding a new feature amount to at least one intermediate layer included in a learned first neural network that has learned a task of performing first class classification, a step of generating a second neural network having a structure in which a network structure of a calculation path of an existing feature amount of the first neural network is maintained and a new feature amount of a subsequent layer is calculated by performing processing of convolving each of the existing feature amount and the new feature amount, and a step of causing the second neural network to acquire a processing function of performing second class classification by performing learning of the second neural network using a set of learning data corresponding to the second class classification.

公开(公告)号：US20230306605A1

公开(公告)日：2023-09-28

申请号：US18170532

申请日：2023-02-17

Applicant: FUJIFILM Corporation

Inventor： Saeko SASUGA ,  Yoshiro Kitamura ,  Akira Kudo

IPC: G06T7/11 ,  G06T11/00 ,  G06T7/174 ,  G06T7/00 ,  G16H30/40

CPC classification number: G06T7/11 ,  G06T11/00 ,  G06T7/174 ,  G06T7/0012 ,  G16H30/40 ,  G06T2207/30008 ,  G06T2207/30096 ,  G06T2207/20081

Abstract: A processor is configured to acquire an original image and a mask image in which masks are applied to one or more regions respectively representing one or more objects including a target object in the original image, derive a pseudo mask image by processing the mask in the mask image, and derive a pseudo image that has a region based on a mask included in the pseudo mask image and has the same representation format as the original image, based on the original image and the pseudo mask image.

公开(公告)号：US11468659B2

公开(公告)日：2022-10-11

申请号：US17000363

申请日：2020-08-23

Applicant: FUJIFILM Corporation

Inventor： Akimichi Ichinose ,  Keigo Nakamura ,  Yoshiro Kitamura

IPC: G06V10/20 ,  G06K9/62 ,  G06N20/00 ,  G06T7/00

Abstract: A learning support device 18 includes an acquisition unit 26, a registration unit 27, a storage device 28, a learning unit 29, and a controller 31. The acquisition unit 26 acquires an image of a region of interest and a name of the region of interest by analyzing an interpretation report 23. The registration unit 27 registers training data consisting of the image of the region of interest and the name of the region of interest acquired by the acquisition unit 26 in the storage device 28. The learning unit 29 performs learning for generating a discrimination model 34, which outputs the image of the region of interest and the name of the region of interest with respect to an input of an inspection image 22 of the interpretation report 23, using a plurality of pieces of training data 33 registered in the storage device 28.

公开(公告)号：US10198669B2

公开(公告)日：2019-02-05

申请号：US15464997

申请日：2017-03-21

Applicant: FUJIFILM Corporation

Inventor： Yoshiro Kitamura

IPC: G06K9/00 ,  G06K9/62 ,  G06T7/00 ,  G06N3/08 ,  G06T7/11

Abstract: [Objective]To enable a three dimensional image to be accurately classified into a plurality of classes with a small amount of calculations, in an image classifying apparatus, an image classifying method, and an image classifying program.[Constitution]A three dimensional image is classified into a plurality of classes by a convolutional neural network, in which a plurality of processing layers are connected hierarchically. The convolutional neural network includes: a convoluting layer that administers a convoluting process on each of a plurality of two dimensional images, which are generated by the neural network administering a projection process on the three dimensional image according to a plurality of processing parameters; and a pooling layer that pools the values of the same position within each of the plurality of two dimensional images which have undergone the convoluting process.

公开(公告)号：US09336458B2

公开(公告)日：2016-05-10

申请号：US14542418

申请日：2014-11-14

Applicant: Fujifilm Corporation

Inventor： Yoshiro Kitamura

IPC: G06K9/62 ,  G06T7/00

CPC classification number: G06K9/6211 ,  G06K2209/051 ,  G06T7/149 ,  G06T7/162 ,  G06T2207/20072 ,  G06T2207/20164 ,  G06T2207/30101 ,  G06T2207/30172

Abstract: An image processing device includes a candidate point extracting unit configured to extract a plurality of candidate points belonging to a predetermined structure from image data, a shape model storing unit configured to store a shape model representing a known shape of the predetermined structure, the shape model being formed by a plurality of model labels having a predetermined connection relationship, and a corresponding point selecting unit configured to select a mapping relationship between the candidate points and the model labels from a set of candidate mapping relationships.

Abstract translation: 一种图像处理装置，包括：候选点提取单元，被配置为从图像数据提取属于预定结构的多个候选点;形状模型存储单元，被配置为存储表示所述预定结构的已知形状的形状模型，所述形状模型 由具有预定连接关系的多个模型标签形成，以及对应点选择单元，被配置为从一组候选映射关系中选择候选点和模型标签之间的映射关系。

公开(公告)号：US09311732B2

公开(公告)日：2016-04-12

申请号：US14668243

申请日：2015-03-25

Applicant: FUJIFILM Corporation

Inventor： Yoshiro Kitamura

IPC: G06K9/00 ,  G06T11/20 ,  G06T7/00

CPC classification number: G06T11/206 ,  G06T7/12 ,  G06T7/162 ,  G06T2207/10004 ,  G06T2207/20072 ,  G06T2207/30101

Abstract: When using a graph cut process for binary labeling, labeling unit selects N (>3) pixels in image data in such a manner to represent a predetermined shape in the image, minimize the high-order energy of the Nth order or greater in which the pixel values of the N pixels are variables, and performs labeling.

Abstract translation: 当使用用于二进制标签的图形切割处理时，标签单元以图像数据中的N（> 3）个像素来选择图像中的预定形状，将第N阶或更高的高阶能量最小化， N像素的像素值是变量，并执行标签。

公开(公告)号：US08515133B2

公开(公告)日：2013-08-20

申请号：US13670678

申请日：2012-11-07

Applicant: Fujifilm Corporation

Inventor： Yoshiro Kitamura

IPC: G06K9/00

CPC classification number: G06K9/469 ,  G06K9/6211 ,  G06K9/6892 ,  G06K2209/05 ,  G06T7/12 ,  G06T7/149 ,  G06T2207/30048

Abstract: A plurality of candidate points are extracted from image data. The plurality of candidate points are normalized, and a set of representative points composing form model that is most similar to set form is selected from the plurality of candidate points. Further, the candidate points and the form model are compared with each other, and correction is performed by adding a region forming structure or by deleting a region, or the like. Accordingly, the structure is detected in image data.

Abstract translation: 从图像数据中提取多个候选点。 多个候选点被归一化，并且从多个候选点中选择构成与设定形式最相似的形式模型的一组代表点。 此外，将候选点和形式模型相互比较，并且通过添加区域形成结构或通过删除区域等来执行校正。 因此，在图像数据中检测结构。