公开(公告)号：US11816832B2

公开(公告)日：2023-11-14

申请号：US16951931

申请日：2020-11-18

Applicant: CANON MEDICAL SYSTEMS CORPORATION

Inventor： Qiulin Tang ,  Jian Zhou ,  Zhou Yu

IPC: G06T7/00 ,  G16H30/40 ,  G16H50/20 ,  G16H50/70 ,  G06N3/08 ,  G06T7/12 ,  G06T11/00 ,  A61B6/03 ,  A61B6/00 ,  G16H50/50

CPC classification number: G06T7/0012 ,  A61B6/032 ,  A61B6/504 ,  A61B6/5264 ,  G06N3/08 ,  G06T7/12 ,  G06T11/005 ,  G16H30/40 ,  G16H50/20 ,  G16H50/50 ,  G16H50/70 ,  G06T2207/10081 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/30101 ,  G06T2211/436

Abstract: Devices, systems, and methods obtain scan data that were generated by scanning a scanned region, wherein the scan data include groups of scan data that were captured at respective angles; generate partial reconstructions of at least a part of the scanned region, wherein each partial reconstruction of the partial reconstructions is generated based on a respective one or more groups of the groups of scan data, and wherein a collective scanning range of the respective one or more groups is less than the angular scanning range; input the partial reconstructions into a machine-learning model, which generates one or more motion-compensated reconstructions of the at least part of the scanned region based on the partial reconstructions; calculate a respective edge entropy of each of the one or more motion-compensated reconstructions of the at least part of the scanned region; and adjust the machine-learning model based on the respective edge entropies.

公开(公告)号：US11786205B2

公开(公告)日：2023-10-17

申请号：US17554032

申请日：2021-12-17

Applicant: CANON MEDICAL SYSTEMS CORPORATION

Inventor： Chung Chan ,  Jian Zhou ,  Evren Asma

IPC: A61B6/03 ,  G06N3/08 ,  G06T7/00 ,  A61B6/00

CPC classification number: A61B6/5258 ,  A61B6/032 ,  A61B6/037 ,  G06N3/08 ,  G06T7/0012 ,  G06T2207/10004 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: A deep learning (DL) convolution neural network (CNN) reduces noise in positron emission tomography (PET) images, and is trained using a range of noise levels for the low-quality images having high noise in the training dataset to produceuniform high-quality images having low noise, independently of the noise level of the input image. The DL-CNN network can be implemented by slicing a three-dimensional (3D) PET image into 2D slices along transaxial, coronal, and sagittal planes, using three separate 2D CNN networks for each respective plane, and averaging the outputs from these three separate 2D CNN networks. Feature-oriented training can be implemented by segmenting each training image into lesion and background regions, and, in the loss function, applying greater weights to voxels in the lesion region. Other medical images (e.g. MRI and CT) can be used to enhance resolution of the PET images and provide partial volume corrections.

公开(公告)号：US20210290191A1

公开(公告)日：2021-09-23

申请号：US16938463

申请日：2020-07-24

Applicant: The Regents of the University of California ,  Canon Medical Systems Corporation

Inventor： Jinyi QI ,  Nimu Yuan ,  Jian Zhou

IPC: A61B6/00 ,  G06N3/08

Abstract: First and second substantially independent identically distributed half scans are obtained; the first substantially independent identically distributed half scan is used as training data to train a machine learning-based system, and the second substantially independent identically distributed half scan is used as label data to train a machine learning-based system. This produces a trained machine learning-based system.

公开(公告)号：US20210007695A1

公开(公告)日：2021-01-14

申请号：US16510632

申请日：2019-07-12

Applicant: Canon Medical Systems Corporation

Inventor： Yujie LU ,  Zhou Yu ,  Jian Zhou

IPC: A61B6/00 ,  A61B6/03 ,  G06T11/00 ,  G06N3/08 ,  G06N20/00

Abstract: A method and apparatus is provided that uses a deep learning (DL) network to improve the image quality of computed tomography (CT) images, which were reconstructed using an analytical reconstruction method. The DL network is trained to use physical-model information in addition to the analytical reconstructed images to generate the improved images. The physical-model information can be generated, e.g., by estimating a gradient of the objective function (or just the data-fidelity term) of a model-based iterative reconstruction (MBIR) method (e.g., by performing one or more iterations of the MBIR method). The MBIR method can incorporate physical models for X-ray scatter, detector resolution/noise/non-linearities, beam-hardening, attenuation, and/or the system geometry. The DL network can be trained using input data comprising images reconstructed using the analytical reconstruction method and target data comprising images reconstructed using the MBIR method.

公开(公告)号：US10803984B2

公开(公告)日：2020-10-13

申请号：US16143161

申请日：2018-09-26

Applicant: Canon Medical Systems Corporation

Inventor： Jian Zhou ,  Zhou Yu ,  Yan Liu

IPC: G16H30/20 ,  G06T5/00 ,  G06T5/50 ,  G06T11/00 ,  G06T7/00 ,  G06K9/62 ,  G06K9/68 ,  G06K9/66 ,  G06N3/08

Abstract: A medical image processing apparatus according to an embodiment comprises a memory and processing circuitry. The memory is configured to store a plurality of neural networks corresponding to a plurality of imaging target sites, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding imaging target site. The processing circuitry is configured to process first data into second data using, among the neural networks, the neural network corresponding to the imaging target site for the first data, wherein the first data is input to the input layer and the second data is output from the output layer.

公开(公告)号：US20200305806A1

公开(公告)日：2020-10-01

申请号：US16370230

申请日：2019-03-29

Applicant: Canon Medical Systems Corporation

Inventor： Qiulin TANG ,  Jian Zhou ,  Zhou Yu

IPC: A61B6/03 ,  G06N3/08 ,  A61B6/00

Abstract: A method and apparatuses are provided that use a neural network to correct artifacts in computed tomography (CT) images, especially cone-beam CT (CBCT) artifacts. The neural network is trained using a training dataset of artifact-minimized images paired with respective artifact-exhibiting images. In some embodiments, the artifact-minimized images are acquired using a small cone angle for the X-ray beam, and the artifact-exhibiting images are acquired either by forwarding projecting the artifact-minimized images using a large-cone-angle CBCT configuration or by performing a CBCT scan. In some embodiments, the network is a 2D convolutional neural network, and an artifact-exhibiting image is applied to the neural network as 2D slices taken for the coronal and/or sagittal views. Then the 2D image results from the neural network are reassembled as a 3D imaging having reduced imaging artifacts.

公开(公告)号：US20190236763A1

公开(公告)日：2019-08-01

申请号：US15884089

申请日：2018-01-30

Applicant: Canon Medical Systems Corporation

Inventor： Chung Chan ,  Zhou Yu ,  Jian Zhou

IPC: G06T5/50 ,  G06T5/00 ,  G06T11/00 ,  A61B6/03 ,  A61B6/00

CPC classification number: G06T5/50 ,  A61B6/032 ,  A61B6/463 ,  A61B6/5235 ,  A61B6/5258 ,  G06T5/002 ,  G06T11/006 ,  G06T11/008 ,  G06T2207/10081 ,  G06T2207/20216 ,  G06T2207/30016 ,  G06T2207/30061 ,  G06T2210/41 ,  G06T2211/424

Abstract: A method and apparatus is provided generate a display image that optimize a tradeoff between resolution and noise by using blending weights/ratio based on the content/context of the image. The blending weights control the relative weights when combining multiple computed tomography (CT) images having different degrees of smoothing/denoising to generate a display image having the optimal tradeoff lying within the continuum between/among the CT images. The blending weights are automated based on information indicating the content/context of the display image (e.g., the segmented tissue type, average attenuation, and the display setting such as window width and window level). Thus, indicia indicating content/context of the image determine the weighting coefficients, which are used in a weighted sum, e.g., to combine the plurality of images with different noise/smoothing parameters into a single blended image, which is displayed.

公开(公告)号：US20240144551A1

公开(公告)日：2024-05-02

申请号：US18049953

申请日：2022-10-26

Applicant: CANON MEDICAL SYSTEMS CORPORATION

Inventor： Yujie Lu ,  Liang Cai ,  Jian Zhou

IPC: G06T11/00 ,  G06T7/00

CPC classification number: G06T11/005 ,  G06T7/0012 ,  G06T2207/10081 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/30008

Abstract: According to some embodiments, a method comprises obtaining a set of projection data acquired from a CT scan of an object; generating, based on the set of projection data, one or more sets of preliminary scattering data; and performing X-ray scatter correction by inputting the obtained set of projection data and the generated one or more sets of preliminary scattering data into a trained machine-learning model for extracting X-ray scatter components from the set of projection data.

公开(公告)号：US11864939B2

公开(公告)日：2024-01-09

申请号：US17339093

申请日：2021-06-04

Applicant: CANON MEDICAL SYSTEMS CORPORATION

Inventor： Jian Zhou ,  Ruoqiao Zhang ,  Zhou Yu ,  Yan Liu

IPC: A61B6/00 ,  G06T11/00 ,  G06T5/50 ,  G06T5/20 ,  G06T5/10 ,  G06N3/084 ,  G06N3/045

CPC classification number: A61B6/5258 ,  A61B6/4014 ,  G06N3/045 ,  G06N3/084 ,  G06T5/10 ,  G06T5/20 ,  G06T5/50 ,  G06T11/005 ,  G06T11/006 ,  G06T11/008 ,  A61B6/482 ,  A61B6/5205 ,  G06T2207/10081 ,  G06T2207/20064 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2211/408

Abstract: A deep learning (DL) network corrects/performs sinogram completion in computed tomography (CT) images based on complementary high- and low-kV projection data generated from a sparse (or fast) kilo-voltage (kV)-switching CT scan. The DL network is trained using inputs and targets, which respectively generated with and without kV switching. Another DL network can be trained to correct sinogram-completion errors in the projection data after a basis/material decomposition. A third DL network can be trained to correct sinogram-completion errors in reconstructed images based on the kV-switching projection data. Performance of the DL network can be improved by dividing a 3D convolutional neural network (CNN) into two steps performed by respective 2D CNNs. Further, the projection data and DLL can be divided into high- and low-frequency components to improve performance.

公开(公告)号：US11801029B2

公开(公告)日：2023-10-31

申请号：US17554019

申请日：2021-12-17

Applicant: CANON MEDICAL SYSTEMS CORPORATION

Inventor： Chung Chan ,  Jian Zhou ,  Evren Asma

IPC: A61B6/00 ,  A61B6/03 ,  G06N3/08 ,  G06T7/00

CPC classification number: A61B6/5258 ,  A61B6/032 ,  A61B6/037 ,  G06N3/08 ,  G06T7/0012 ,  G06T2207/10004 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: A deep learning (DL) convolution neural network (CNN) reduces noise in positron emission tomography (PET) images, and is trained using a range of noise levels for the low-quality images having high noise in the training dataset to produce uniform high-quality images having low noise, independently of the noise level of the input image. The DL-CNN network can be implemented by slicing a three-dimensional (3D) PET image into 2D slices along transaxial, coronal, and sagittal planes, using three separate 2D CNN networks for each respective plane, and averaging the outputs from these three separate 2D CNN networks. Feature-oriented training can be implemented by segmenting each training image into lesion and background regions, and, in the loss function, applying greater weights to voxels in the lesion region. Other medical images (e.g. MRI and CT) can be used to enhance resolution of the PET images and provide partial volume corrections.