Shimadzu Corporation

Murata Manufacturing Co., Ltd.

Wataru Takahashi

Shinichiro Mori

Michel Dargis

Shota Oshikawa

Ayako Akazawa

Keiichi Goto

Takanori Yoshida

Yusuke Tagawa

Toshiki Matsui

Takihito Sakai

Shota Sato

Katsuya Ikegami

Hidetaka Takezawa

Frederic Hudon

Kodai Nagae

Yuki Ohmae

Yohei Sugawara

Sebastien Matte

Zhengyan Gao

Kazue Mizuno

Yuichiro Hirano

Jun Sakatsume

In a creation method of a trained model, a reconstructed image () obtained by reconstructing three-dimensional X-ray image data () is generated. A projection image () is generated from a three-dimensional model of an image element () by a simulation. The projection image is superimposed on the reconstructed image to generate a superimposed image (). A trained model () is created by performing machine learning using the superimposed image, and the reconstructed image or the projection image.

Creation method of trained model, image generation method, and image processing device

A cell image analysis method includes converting a first image into a label image by performing a segmentation process to identify a region of a cell that has already started differentiation and a region of an undifferentiated cell in the first image, acquiring a shape feature amount from the label image, and determining whether or not a cell colony includes a colony region that is a candidate for a search target based on the shape feature amount and a determination criterion.

Cell image analysis method and cell image analyzer

This image generation device is provided with: a learning image generation unit () for generating a training input image and a training output image based on three-dimensional data; a noise addition unit () for adding the same noise to the training input image and the training output image; a learning unit () for learning a learning model for extracting or removing a specific portion by performing machine learning based on the training input image to which the noise has been added and the training output image to which the noise has been added; and an image generation unit () for generating an image from which the specific portion has been extracted or removed by using a learned learning model.

Image generation device, image generation method, and learned model generation method

A filter device includes a multilayer substrate, an inductor, and a first open stub. The multilayer substrate includes a first wiring layer, a second wiring layer including a first reference electrode to which a reference potential is supplied, and a plurality of first dielectric layers between the first wiring layer and the second wiring layer. The inductor has one end coupled to a first terminal and another end coupled to a second terminal. A portion of the first open stub is provided in the first wiring layer, and the first open stub has one end coupled to the one end of the inductor and another open-circuited end. When the multilayer substrate is viewed in plan view in a stacking direction, a capacitor is formed by the first reference electrode and the first open stub that mutually overlap. A resonant circuit is formed by the capacitor and the inductor.

Filter device

A variable circuit includes a switch including a plurality of input terminals and a plurality of output terminals and an external wiring line. The multiple input terminals include a first input terminal to which a first input signal is inputted and a second input terminal to which a second input signal is inputted. The multiple output terminals include a first output terminal from which a first output signal is outputted and a second output terminal from which a second output signal is outputted. The switch is capable of forming at least one internal connection path electrically connecting any one of the multiple input terminals and any one of the multiple output terminals. The external wiring line is disposed outside the switch and is configured to electrically connect the second output terminal to the second input terminal.

Variable circuit

A filter device includes: a multilayer board including a first dielectric layer, a first wiring layer including a reference electrode to which a reference potential is supplied, and a second dielectric layer positioned between the first dielectric layer and the first wiring layer and having a thickness different from a thickness of the first dielectric layer; a series passive element provided on a series wire electrically connecting a first terminal and a second terminal together; a first parallel wire electrically connecting the series wire and the reference electrode together; and a first parallel passive element provided on the first parallel wire. The first parallel wire includes: a first via penetrating through the first dielectric layer and electrically connected to the first parallel passive element; and a second via penetrating through the second dielectric layer and electrically connecting the first via and the reference electrode together.

This production method of a trained model () includes a step of optimizing a training model () so that a value of a loss function (A) becomes small. The loss function (A) is configured to output a relatively large value in a case where the contrast of the predetermined area (PA) in a brightness-adjusted image adjusted in brightness becomes high with respect to training data () than in a case where the contrast of the predetermined area in the brightness-adjusted image () becomes low with respect to the training data.

Production method of learned model, brightness adjustment method, and image processing apparatus

A phase image is formed by calculation from a hologram image of a cell, and segmentation is performed for each pixel for the phase image using a fully convolution neural network to identify an undifferentiated cell region, a deviated cell region, a foreign substance region, and the like. When learning, when a learning image included in a mini-batch is read, the image is randomly inverted vertically or horizontally and then is rotated by a random angle. A part that has been lost within the frame by the pre-rotation image is compensated for by a mirror-image inversion with an edge of a post-rotation image as an axis thereof. Learning of a fully convolution neural network is performed using the generated learning image. The same processing is repeated for all mini-batches, and the learning is repeated by a predetermined number of times while shuffling the training data allocated to the mini-batch. The precision of the learning model is thus improved. In addition, since rotationally invariant characteristics can be learned, it is possible to identify cell colonies of various shapes with good precision.

Cellular image analysis method, cellular image analysis device, and learning model creation method

An image generating device for generating an image which is an X-ray image of an area including a bone portion of a subject with the bone portion removed has a control unitincluding: a DRR imagerconfigured to generate a first DRR image of an area including a bone portion and a second DRR image showing the bone portion, by performing, for a set of CT image data of an area including the bone portion of a subject, a virtual fluoroscopic projection simulating a geometric fluoroscopy condition of an X-ray irradiator and an X-ray detector for the subject; a training sectionconfigured to generate a machine learning model for recognizing the bone portion, by performing machine learning using the first DRR image and the second DRR image serving as a training image; an image converterconfigured to perform conversion of the X-ray image of the area including the bone portion of the subject, using the machine learning model trained in the training section, to generate an image showing the bone portion; and a bone portion subtractorconfigured to subtract the image showing the bone portion from the X-ray image of the area including the bone portion of the subject.

Image creation device

A training label image correction method includes performing a segmentation process on an input image () of training data () by a trained model () using the training data to create a determination label image (), comparing labels of corresponding portions in the determination label image () and a training label image () with each other, and correcting label areas () included in the training label image based on label comparison results.

Training label image correction method, trained model creation method, and image analysis device

A cell image analysis apparatus that can achieve less time and effort for labeling for generation of teaching data than in a conventional example is provided. The cell image analysis apparatus includes an image obtaining unit that obtains a cell image including a removal target that is obtained by a microscope for observation of a cell, a teaching data generator that specifies a removal target region including the removal target within the cell image by performing predetermined image processing and generates as teaching data for machine learning, a label image that represents a location of the removal target region in the cell image, and a training data set generator that generates a set of the cell image and the label image as a training data set to be used in machine learning.

Cell image analysis apparatus, cell image analysis system, method of generating training data, method of generating trained model, training data generation program, and method of producing training data

This radiographic imaging apparatus () is provided with a control unit () configured to associate a projection image () in which at least a part of a post-removal projection image () in which a contrast agent image (Ba) has been removed and at least a part of a fluoroscopic image () are most similar.

Radiographic imaging apparatus

A control partincludes: a DRR image creation partthat creates a DRR image including a specific site; a specific site projection partthat creates a projection region image representing the region of the specific site; a discriminator learning partthat learns a discriminator for recognizing the region of the specific site by performing machine learning with use of the DRR image and the projection region image as a training label image; a specific site region detection partthat detects the region of the specific site by performing discrimination using the discriminator learned by the discriminator learning parton an X-ray fluoroscopic image including the specific site; and a radiation signal generation partthat transmits a treatment beam radiation signal to an irradiation devicewhen the region of the specific site detected by the specific site region detection partis included in the radiation region of a treatment beam.

Tracking device for radiation treatment, position detection device, and method for tracking moving body

A full-size training image is reduced by an image reduction unit () and input to an FCN (Fully Convolutional Neural Network) computation unit (), and the FCN computation unit () performs calculation under a set filter coefficient and outputs a reduced label image. The reduced label image is enlarged to a full size by an image enlargement unit (), the error calculation unit () calculates an error between the enlarged label image and a full-size ground truth based on a loss function, and the parameter update unit () updates a filter coefficient depending on the error. By repeating learning under the control of the learning control unit, it is possible to generate a learning model for performing optimal segmentation including an error generated at the time of image enlargement. Further, by including the image enlargement processing in the learning model, a full-size label image can be output, and the accuracy evaluation of the model can also be performed with high accuracy.

Machine learning method, machine learning device, and machine learning program

A control unitincludes: an image storage unitconstituted by a first image storage unitthat stores multiple templates created on the basis of an image including a specific site of a subject and a second image storage unitthat stores multiple positive images created on the basis of an image including the specific site of the subject; a learning unitthat, on the basis of the multiple positive images, creates a discriminator by machine learning; a position selection unitthat, with use of multiple images obtained by collecting an image including the specific site of the subject at a predetermined frame rate, selects a region including the specific site by machine learning using the discriminator; and a position detection unitthat detects the position of the specific site by performing template matching using the multiple templates on the region including the specific site selected by the position selection unit

Growing community of inventors

Kyoto, Japan

Wataru Takahashi