公开(公告)号：US11830111B2

公开(公告)日：2023-11-28

申请号：US17047062

申请日：2018-05-17

Applicant: Shenzhen Institutes of Advanced Technology

Inventor： Yanjie Zhu ,  Yuanyuan Liu ,  Dong Liang ,  Xin Liu ,  Hairong Zheng

IPC: G06T11/00 ,  G01R33/56 ,  G01R33/561 ,  A61B5/055

CPC classification number: G06T11/005 ,  G01R33/561 ,  G01R33/5608 ,  G06T11/006 ,  A61B5/055

Abstract: Provided are a method and device for magnetic resonance parameter imaging, medical equipment, and a storage medium. The method comprises: performing acceleration sampling with respect to an image to be reconstructed of an observation target in a preset parameter direction to acquire K-space data corresponding to the image to be reconstructed, calculating, on the basis of the K-space data and of a parameter relaxation model, a parameter value and a compensation coefficient of the image to be reconstructed, generating, on the basis of the compensation coefficient, a compensation image corresponding to the image to be reconstructed, calculating, on the basis of the compensation image, respectively a low-rank part and a sparse part of the image to be reconstructed so as to update the compensation image.

公开(公告)号：US11756161B2

公开(公告)日：2023-09-12

申请号：US17340117

申请日：2021-06-07

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY

Inventor： Zhanli Hu ,  Hairong Zheng ,  Na Zhang ,  Xin Liu ,  Dong Liang ,  Yongfeng Yang ,  Hanyu Sun

IPC: G06T5/00 ,  A61B6/03 ,  G06N3/08 ,  G06N3/045

CPC classification number: G06T5/00 ,  A61B6/037 ,  G06N3/045 ,  G06N3/08 ,  G06T2207/10088 ,  G06T2207/10104 ,  G06T2207/20081

Abstract: The present application relates to a method and system for generating multi-task learning-type generative adversarial network for low-dose PET reconstruction, and relates to the field of deep learning. The method includes connecting layers of the encoder with layers of the decoder by skip connection to provide a U-Net type picture generator; generating a group of generative adversarial networks by matching a plurality of picture generators with a plurality of discriminators in one-to-one manner; obtaining a first multi-task learning-type generative adversarial network; designing a joint loss function 1 for improving image quality; and training the first multi-task learning-type generative adversarial network according to the joint loss function 1 in combination with an optimizer to provide a second multi-task learning-type generative adversarial network.

公开(公告)号：US11508048B2

公开(公告)日：2022-11-22

申请号：US16976474

申请日：2020-02-10

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY

Inventor： Zhan Li Hu ,  Dong Liang ,  Yong Chang Li ,  Hai Rong Zheng ,  Yong Feng Yang ,  Xin Liu

IPC: G06T5/50 ,  G06T5/00

Abstract: The present disclosure discloses a method and a system for generating a composite PET-CT image based on a non-attenuation-corrected PET image. The method includes: constructing a first generative adversarial network and a second generative adversarial network; obtaining a mapping relationship between a non-attenuation-corrected PET image and an attenuation-corrected PET image by training the first generative adversarial network; obtaining a mapping relationship between the attenuation-corrected PET image and a CT image by training the second generative adversarial network; and generating the composite PET-CT image by utilizing the obtained mapping relationships. According to the present disclosure, a high-quality PET-CT image can be directly composited from a non-attenuation-corrected PET image, and medical costs can be reduced for patients, and radiation doses applied to the patients in examination processes can be minimized.

公开(公告)号：US11327137B2

公开(公告)日：2022-05-10

申请号：US16761285

申请日：2017-06-06

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY

Inventor： Shanshan Wang ,  Dong Liang ,  Ningbo Huang ,  Xin Liu ,  Hairong Zheng

IPC: G01R33/561 ,  G01R33/00 ,  G06N3/08

Abstract: The present disclosure relates to a 1D partial Fourier parallel magnetic resonance imaging method with a deep convolutional network and belongs to the technical field of magnetic resonance imaging. The method includes steps of: creating a sample set and a sample label set for training; constructing an initial deep convolutional network model; inputting a training sample of the sample set to the initial deep convolutional network model for forward process, comparing an output result of the forward process with an expected result in the sample label set, and performing training with a gradient descent method until a parameter of each layer which enables consistency between the output result and the expected result to be maximum is obtained; creating an optimal deep convolutional network model by using the obtained parameter of the each layer; and inputting a multi-coil undersampled image sampled online to the optimal deep convolutional network model, performing the forward process on the optimal deep convolutional network model, and outputting a reconstructed single-channel full-sampled image. The present disclosure can well remove the noise of the reconstructed image, reconstruct a magnetic resonance image with a better visual effect, and has high practical value.

公开(公告)号：US11125843B2

公开(公告)日：2021-09-21

申请号：US16309351

申请日：2017-12-29

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY

Inventor： Yin Wu ,  Hairong Zheng ,  Xin Liu

IPC: G01R33/56

Abstract: A method for measuring pH includes using amide protons as an endogenous contrast agent, measuring a chemical exchange saturation transfer effect ratio R of amide protons corresponding to a pH-known amide proton solution under different saturation powers, then establishing a function relation between pH and R according to different pH-known amide proton solutions and the corresponding measured R thereto, and finally calculating the desired pH according to experimentally measured Ri and the function relation. The method can eliminate the impact of concentration and does not require any estimation or measurement of the parameters such as the concentration of exchangeable protons, the longitudinal relaxation time of water, etc., and therefore can measure pH more accurately, conveniently and non-invasively.

公开(公告)号：US11668777B2

公开(公告)日：2023-06-06

申请号：US16760956

申请日：2017-12-01

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY

Inventor： Shanshan Wang ,  Dong Liang ,  Sha Tan ,  Xin Liu ,  Hairong Zheng

IPC: G01R33/561 ,  G01R33/56 ,  G06T11/00

CPC classification number: G01R33/5611 ,  G01R33/5608 ,  G06T11/008 ,  G06T2207/10088 ,  G06T2207/20056 ,  G06T2207/20081

Abstract: Provided are a parallel magnetic resonance imaging method and apparatus based on adaptive joint sparse codes and a computer-readable medium. The method includes solving an l2−lF−l2,1 minimization objective, where the l2 norm is a data fitting term, the lF norm is a sparse representation error, and the l2,1 mixed norm is the joint sparsity constraining across multiple channels; separately updating each of a sparse matrix, a dictionary and K-space data with a corresponding algorithm, and obtaining a reconstructed image by a sum of root mean squares of all the channels. The joint sparsity of the channels is developed using the norm l2,1. In this manner, calibration is not required while information sparsity is developed. Moreover, the method is robust.

公开(公告)号：US11579229B2

公开(公告)日：2023-02-14

申请号：US16473806

申请日：2018-07-24

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY

Inventor： Haifeng Wang ,  Dong Liang ,  Sen Jia ,  Xin Liu ,  Hairong Zheng

IPC: G01V3/00 ,  G01R33/561 ,  G01R33/56 ,  G01R33/58

Abstract: There are provided a parallel rapid imaging method and device based on a complex number conjugate symmetry of multi-channel coil data and nonlinear GRAPPA image reconstruction, and a medium. The imaging method includes: obtaining virtual conjugate coil data by expanding the actual multi-channel coil data; combining actual multi-channel coil data and virtual multi-channel coil data to obtain a linear data term and a nonlinear data term; calibrating weighting factors of the linear data term and the nonlinear data term by using combined low-frequency full-sampling data (margins of the low-frequency full-sampling data includes parts of high-frequency data); reconstructing data which is under-sampled in a high-frequency region according to the calibrated weighting factors; fusing the low-frequency full-sampling data and the reconstructed data for the high-frequency region.

公开(公告)号：US11514621B2

公开(公告)日：2022-11-29

申请号：US16878633

申请日：2020-05-20

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY

Inventor： Zhan Li Hu ,  Dong Liang ,  Hai Rong Zheng ,  Xin Liu ,  Yong Feng Yang ,  Zhen Xing Huang

IPC: G06T11/00 ,  A61N5/10

Abstract: The disclosure provides a low-dose image reconstruction method and system based on prior anatomical structure difference. The method includes: determining the weights of different parts in the low-dose image based on prior information of anatomical structure differences; constructing a generative network being taking the low-dose image as input extract features, and integrating the weights of the different parts in the feature extraction process, outputting a predicted image; constructing a determining network being taking the predicted image and standard-dose image as input, to distinguish the authenticity of the predicted image and standard-dose image as the first optimization goal, and identifying different parts of the predicted image as the second optimization goal, collaboratively training the generative network and the determining network to obtain the mapping relationship between the low-dose image and the standard-dose image; and reconstructing the low-dose image by using the obtained mapping relationship. The disclosure can obtain more accurate high-definition images.

公开(公告)号：US10782379B2

公开(公告)日：2020-09-22

申请号：US16645798

申请日：2015-12-30

Applicant: SHENZHEN INSTITUTES OF ADVANCED TECHNOLOGY CHINESE ACADEMY OF SCIENCES

Inventor： Hairong Zheng ,  Xin Liu ,  Chuanli Cheng ,  Chao Zou

IPC: G01R33/565 ,  A61B5/055

Abstract: A method, an apparatus, and a device for magnetic resonance chemical shift encoding imaging are provided. The method includes in a phasor-error spectrum established based on a simplified multi-point magnetic resonance signal model, determining a pixel point having a unique phase factor value and causing the phasor-error spectrum to reach a local minimum value as an initial seed point; estimating a phase factor value of a pixel point to be estimated according to the initial seed point to obtain a field map; mapping and merging the field maps at the highest resolution to obtain a reconstructed field map; determining a reconstructed seed point from the reconstructed field map, and estimating the reconstructed seed point to obtain a phase factor value of a reconstructed pixel point to be estimated.

公开(公告)号：US20190011515A1

公开(公告)日：2019-01-10

申请号：US16067204

申请日：2015-12-30

Applicant: Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences

Inventor： Hairong Zheng ,  Xin Liu ,  Chuanli Cheng ,  Chao Zou

IPC: G01R33/48 ,  G01R33/565 ,  G01R33/485 ,  G01R33/46 ,  A61B5/055 ,  G01R33/44 ,  G01R33/56

Abstract: Provided are a magnetic resonance chemical-shift-encoded imaging method, apparatus, and device, belonging to the technical field of magnetic resonance imaging. The method comprises: in a phasor-error plot established on the basis of a two-point magnetic resonance signal model, determining to be an initial seed point a pixel having a unique phasor and causing said plot to reach a minimal local value; according to the initial seed point, estimating the phasor value of a to-be-estimated pixel to obtain a field map; mapping and merging the field map at the highest resolution to obtain a reconstructed field map; determining a reconstructed seed point from the reconstructed field map, and estimating the reconstructed seed point to obtain the phasor value of the reconstructed to-be-estimated pixel; according to the reconstructed seed point and the phasor value of the reconstructed to-be-estimated pixel, obtaining two separate images having predetermined components. In the method, a region simultaneously containing two components is identified as a seed point, eliminating the deviation caused by phasor-value jump at high resolution and ensuring the correctness of the seed point ultimately selected.