22 results about "MRI diffusion" patented technology

Disclosed is an unsupervised deep learning method, which simultaneously performs registration between diffusion weighted magnetic resonance images and quantification of diffusion parameters. The unsupervised deep learning method decreases a registration error caused by a contrast difference by comparing a similarity between images of the same contrast, increases the performance of registration, and increases the accuracy of quantifying magnetic resonance diffusion parameters.

The invention relates to a magnetic resonance diffusion microstructure imaging method based on tensor learning. The method comprises the following steps: acquiring dense sampling data of diffusion magnetic resonance imaging; fitting is carried out through a diffusion model, and a plurality of real microstructure parameters are generated; generating sparse sampling data based on the dense sampling data; constructing a deep neural network model, inputting sparse sampling data, and outputting a plurality of predicted microstructure parameters; respectively performing tensor decomposition on the real microstructure parameter and the predicted microstructure parameter to obtain a real singular value tensor and a predicted singular value tensor containing high-order correlation; constructing a joint training function, and training a deep neural network model; and inputting real sparse sampling data into the trained deep neural network model. By means of the arrangement, high-dimensional information of the multi-parameter tensor is reserved, correlation of different microstructure parameters is mined through tensor decomposition, and rapid and accurate microstructure imaging is assisted.

The invention discloses a rapid magnetic resonance diffusion kurtosis parameter imaging method based on a model, and relates to the technical field of magnetic resonance imaging. The method comprises the following steps: performing signal acquisition on a measured target; establishing a DKI mathematical model; calculating sensitivity information S of the multi-channel coil; performing initial reconstruction on the down-sampling k space data of each b value and each diffusion coding direction to obtain an initial complex image; performing DKI tensor estimation according to the amplitude information of the initial complex image to obtain an initial DKI tensor; extracting phase information P of each b value and each diffusion coding direction; constructing a target function with regularization constraint; solving a DKI tensor enabling the objective function to be minimum; calculating a DKI quantization parameter; and calculating a DKI image according to the DKI tensor obtained by optimization solution and the DKI mathematical model. According to the method, the DKI tensor can be directly reconstructed from the downsampling k space data, two steps of image reconstruction and tensor fitting are not needed, error transmission is avoided, and the DKI parameter imaging quality is further improved.

A computer system that performs pairwise harmonization is described. During operation, the computer system may receive information specifying magnetic-resonance diffusion measurements, where the magnetic-resonance diffusion measurements are associated with at least an individual and a first site. Then, the computer system may perform the pairwise harmonizing of the magnetic-resonance diffusion measurements based at least in part on target magnetic-resonance diffusion measurements, where the target magnetic-resonance diffusion measurements are associated with a population and one or more target sites, and the magnetic-resonance diffusion measurements and the target magnetic-resonance diffusion measurements may be acquired using different magnetic-resonance scanners. Next, the computer system may perform quality control on the pairwise harmonized magnetic-resonance diffusion measurements, where the quality control may include comparing the pairwise harmonized magnetic-resonance diffusion measurements associated with the first site and the one or more target sites.

An echo time-dependent magnetic resonance diffusion imaging signal generation method and apparatus. The method comprises: acquiring magnetic resonance non-diffusion-weighted images acquired using a plurality of echo times and magnetic resonance multiple-b-value diffusion-weighted images acquired using a single echo time (S110); on the basis of the non-diffusion-weighted images and quantitative T2 values of multiple water components in human tissues, calculating voxel-by-voxel non-T2 value-weighted magnetic resonance diffusion signals corresponding to the multiple water components in the human tissues (S120); on the basis of the multi-b-value diffusion-weighted images, the magnetic resonance diffusion signals, and the quantitative T2 values, calculating voxel-by-voxel diffusion coefficients corresponding to the multiple water components in the human tissues (S130); and, on the basis of the multi-b-value diffusion-weighted images, the magnetic resonance diffusion signals, the quantitative T2 values and the diffusion coefficients, calculating voxel-by-voxel magnetic resonance diffusion imaging signals acquired using any echo time (S140).

The present invention relates to a low-field magnetic resonance diffusion imaging reconstruction method, comprising the steps of: acquiring low-field magnetic resonance K-space data; mapping the acquired K-space data to a standard Cartesian coordinate system using a Gaussian kernel grid interpolation method, weighting each radial sampling point using a Gaussian kernel function to ensure that the data is smoothly distributed on a standard grid; utilizing the interpolated K-space data to reconstruct the image based on image prior constraints, wherein the image prior constraints include: the image's low-rank structure and sparse prior information; and gradually optimizing the parameters of the low-rank and sparse components using a hybrid kernel function, and combining these constraints with the image's low-rank structure and sparse prior information to enhance the image reconstruction effect. The present invention also relates to a low-field magnetic resonance diffusion imaging reconstruction system. The present invention can significantly improve image quality and noise immunity.

The present invention provides a fast multi-contrast magnetic resonance diffusion imaging and reconstruction method and system, including: using a keyhole echo plane sequence to undersample k-space data of multiple diffusion-weighted images through multiple receiving coils; constructing an objective function including a multi-coil image fidelity term and an image regularization term based on the coil-space-contrast domain tensor low-rank (Low-RankTensorConstraint); minimizing the objective function through an optimization algorithm, reconstructing the diffusion images of all receiving coils and obtaining relevant quantitative parameter images according to the specific way of contrast change. The present invention implements undersampling of k-space through a keyhole-EPI sequence, improves the resolution of multi-contrast diffusion imaging, shortens the echo time, thereby improving the deformation artifacts of the image and reducing the imaging time; avoids the step of solving the coil sensitivity map, reduces the image artifacts caused by variable-density EPI imaging, and improves the robustness and accuracy of the reconstruction result.

A physical intelligent high-definition magnetic resonance diffusion imaging method includes the following steps: 1) Obtain multi-channel single-shot or multi-shot magnetic resonance images with a phase and a b-value of 0 mm / s², and estimate the channel sensitivity; 2) Obtain the simulated motion phase according to the polynomial phase model of rigid body motion; 3) Use the magnetic resonance images with a phase, the channel sensitivity, and the simulated motion phase to generate a large number of multi-shot diffusion-weighted image data as the training data of the intelligent reconstruction network; 4) Construct an intelligent reconstruction network including multiple iterative blocks, where each iterative block includes a low-rank module, a sparse module, and a data verification module. The last iterative block further includes a denoising module. And use the simulated data to train the intelligent reconstruction network; 5) Obtain the multi-shot diffusion-weighted data to be reconstructed; 6) Use the trained intelligent reconstruction network to reconstruct the multi-shot diffusion-weighted data to obtain the reconstructed image.

The application discloses a physical basis model multi-position high-definition diffusion magnetic resonance imaging method, and relates to a multi-shot planar echo diffusion magnetic resonance imaging method. The method comprises the following steps: 1) acquiring k-space data of a diffusion magnetic resonance imaging to be reconstructed, which is collected by a multi-shot planar echo sequence; 2) designing a bidirectional one-dimensional low-rank reconstruction model based on a local smooth phase; 3) generating data by using a physical diffusion model; 4) training a basis model neural network by using the physical generated data; 5) solving the bidirectional one-dimensional low-rank reconstruction model; and 6) combining reconstruction parameters prompted by the basis model neural network, and generalizing reconstruction to obtain multiple-position high-definition diffusion magnetic resonance images without motion artifacts. The method realizes robust high-definition magnetic resonance diffusion imaging of multiple challenging positions without adjusting parameters by using a bidirectional one-dimensional low-rank optimization reconstruction method and a neural network to adaptively optimize reconstruction parameters.

The application provides a magnetic resonance diffusion weighted image denoising method, system, computer device and medium, and belongs to the technical field of image processing. The method utilizes a twice-moving grid division strategy, calculates a standard deviation matrix according to the texture characteristics of the magnetic resonance diffusion weighted image, divides the image into a flat region, a low-texture richness region and a high-texture richness region, and generates a corresponding adaptive mask matrix. In the training stage, the image is masked with a high mask rate; in the inference stage, a low mask rate is used, and the average result of multiple mask inferences is combined to improve the denoising stability. Through the denoising model constructed based on the blind spot network idea, the pixel values of the masked regions are predicted by using the pixels of the unmasked regions, and the model parameters are optimized by using an L2 loss function. The application adopts the above method, improves the definition of the image and retains the texture details, and is not sensitive to the direction number, and stable and good denoising effect can also be achieved when the direction number is small.

The invention discloses a liver cancer iconography diagnosis method based on artificial intelligence, which integrates anatomical structure characteristics of CT enhanced scanning, soft tissue heterogeneity characteristics of MRI diffusion weighted imaging and dynamic blood flow characteristics of ultrasound contrast, and covers three dimensions of morphology, histology and hemodynamics of a focus. The one-sidedness defect of single-mode information is overcome, multi-mode data complementation is achieved, the misdiagnosis and missed diagnosis risks caused by feature missing are effectively reduced, and the diagnosis accuracy is improved. The cross-modal attention module calculates the cosine similarity between each modal and the global reference feature and dynamically distributes the weight, thereby avoiding the limitation of the fixed weight, enabling different modal features to make accurate contributions according to the actual expression of the focus, strengthening the characterization capability of the combined feature for the liver cancer focus, and improving the accuracy of the liver cancer focus. And the recognition precision of the model on difficult focuses such as fuzzy boundaries and complex blood supply is improved.

The invention discloses a transcranial electrical stimulation finite element simulation method and a related device, and the method comprises the steps: carrying out the high-frequency conductivity imaging based on magnetic resonance, and obtaining the high-frequency conductivity; according to the high-frequency conductivity, reconstructing a low-frequency conductivity tensor based on a magnetic resonance diffusion tensor imaging technology; constructing an individual head finite element model with voxel-level conductivity according to the reconstructed low-frequency conductivity tensor; and performing electric field solution on the individual head finite element model by using a finite element method to obtain an electrical stimulation electric field distribution simulation result under the specified electrode. According to the method, the limitation of brain tissue segmentation errors and isotropic hypothesis is avoided, the accuracy and physiological credibility of conductivity modeling and simulation are remarkably improved, and a reliable basis can be provided for individualized optimization design and clinical application of transcranial electrical stimulation.

A physical intelligent magnetic resonance diffusion weighted image denoising and quantification method relates to diffusion magnetic resonance weighted imaging. The method comprises the following steps: 1) obtaining diffusion weighted magnetic resonance data acquired by a planar echo imaging sequence in multiple diffusion directions; 2) designing a deep learning network model and loss function for physical intelligent magnetic resonance diffusion weighted image denoising and quantification; 3) using the training data set obtained in step 1), training a deep learning network model for physical intelligent magnetic resonance diffusion weighted image denoising and quantification; 4) inputting the diffusion weighted data to be denoised into the trained network to generate the final denoising and quantification results. Through the joint application of physical regularization loss and low b-value data, the low signal-to-noise ratio problem in magnetic resonance diffusion weighted imaging is effectively solved, image denoising is achieved, and accurate quantitative results are further provided. While reducing the number of acquisitions, the image quality and ADC quantitative accuracy can be maintained or even improved.

A computer system that performs pairwise harmonization is described. During operation, the computer system may receive information specifying magnetic-resonance diffusion measurements, where the magnetic-resonance diffusion measurements are associated with at least an individual and a first site. Then, the computer system may perform the pairwise harmonizing of the magnetic-resonance diffusion measurements based at least in part on target magnetic-resonance diffusion measurements, where the target magnetic-resonance diffusion measurements are associated with a population and one or more target sites, and the magnetic-resonance diffusion measurements and the target magnetic-resonance diffusion measurements may be acquired using different magnetic-resonance scanners. Next, the computer system may perform quality control on the pairwise harmonized magnetic-resonance diffusion measurements, where the quality control may include comparing the pairwise harmonized magnetic-resonance diffusion measurements associated with the first site and the one or more target sites.

The present disclosure relates to a magnetic resonance diffusion imaging fat suppression method based on echo navigation water-fat separation, comprising: acquiring echo and / or navigation echo corresponding to multiple chemical shift encoding points under the condition that the b value is equal to 0 and the condition that the b value is greater than 0; solving the water map, the fat map, the main magnetic field inhomogeneity and the relaxation time of the target object under the condition that the b value is equal to 0, and determining the water map and the fat map of the target object under the condition that the b value is greater than 0. Through the present disclosure, a chemical shift encoding fat suppression method suitable for clinical use is provided for magnetic resonance EPI diffusion imaging, the interference of fat chemical shift artifacts is reduced, the accuracy of image contrast and quantitative parameter measurement is improved, the universality of chemical shift encoding EPI DWI is improved, and the image reconstruction speed and robustness are accelerated.

The application discloses a kind of based on clustering's magnetic resonance diffusion ordering spectrum processing method, device and readable medium, by obtaining the attenuation spectrum corresponding to several gradient coding magnetic resonance diffusion ordering spectrum, selecting the signal point of each spectral peak on the chemical shift dimension in attenuation spectrum, and constructing the feature vector corresponding to each signal point, length is the number of gradients;According to the feature vector of all signal points of each spectral peak, classification is carried out by clustering algorithm, and the classification result of the signal point of each spectral peak is obtained;According to the classification result of the signal point of each spectral peak and the spectral intensity corresponding to each signal point, statistical analysis is carried out, and the class of each spectral peak is determined, according to the class, the magnetic resonance diffusion ordering spectrum is separated into components, and the separated spectrum corresponding to different components is obtained.The method is simple in operation, widely applicable, excellent in effect, suitable for a small amount of gradient, can effectively shorten time, good in robustness, and small error.

The present invention discloses a model-based rapid magnetic resonance diffusion kurtosis parameter imaging method, which relates to the field of magnetic resonance imaging technology. The method includes: acquiring signals from a target; establishing a DKI mathematical model; calculating sensitivity information S of a multi-channel coil; performing initial reconstruction on downsampled k-space data for each b-value and each diffusion coding direction to obtain an initial complex image; estimating a DKI tensor based on the amplitude information of the initial complex image to obtain an initial DKI tensor; extracting phase information P for each b-value and each diffusion coding direction; constructing an objective function with regularization constraints; solving the DKI tensor that minimizes the objective function; calculating DKI quantization parameters; and calculating a DKI image based on the DKI tensor obtained through optimization and the DKI mathematical model. The present invention can directly reconstruct a DKI tensor from downsampled k-space data without the two steps of image reconstruction and tensor fitting, thus avoiding error propagation and improving the quality of DKI parameter imaging.

The invention discloses a physical basis model multi-part high-definition diffusion magnetic resonance imaging method, and relates to a multi-excitation plane echo diffusion magnetic resonance imaging method. Comprising the following steps: 1) acquiring k space data of diffusion magnetic resonance imaging to be reconstructed acquired by a multi-excitation plane echo sequence; 2) designing a bidirectional one-dimensional low-rank reconstruction model based on a local smooth phase; 3) generating data by using a physical diffusion model; 4) training the base model neural network by using the physical generation data; 5) solving a bidirectional one-dimensional low-rank reconstruction model; and 6) in combination with reconstruction parameters prompted by the basis model neural network, generalization reconstruction is carried out to obtain high-definition diffusion magnetic resonance images of various parts without motion artifacts. According to the method, robust high-definition magnetic resonance diffusion imaging of a plurality of challenge parts without parameter adjustment is realized through a bidirectional one-dimensional low-rank optimization reconstruction method and neural network adaptive optimization reconstruction parameters.

The invention relates to a fat pressing method and device for magnetic resonance diffusion imaging based on water-fat separation, and the method comprises the steps: adjusting the bandwidth corresponding to a target excitation pulse and a target refocusing pulse, so as to change the position of a main peak fat layer excited and refocused in initial magnetic resonance imaging, and inhibit the main peak fat; moving a preset signal reading window to collect diffusion and non-diffusion imaging signals of single excitation or multiple excitation corresponding to the signal reading window at different moving time; constructing a corresponding diffusion imaging signal model, simplifying the diffusion imaging signal model, solving the non-diffusion imaging signal model through a water-fat separation algorithm to obtain a water graph, a fat graph, a magnetic field non-uniformity graph and a relaxation time graph corresponding to the non-diffusion imaging signal model, and calculating a coil sensitivity graph corresponding to the non-diffusion imaging signal model; and solving a diffusion imaging signal model to obtain a diffusion-weighted water map, a fat map and a phase difference between excitation so as to finish inhibition of olefin peak fat.

The invention relates to a low-field magnetic resonance diffusion imaging reconstruction method. The method comprises the following steps: acquiring K space data of low-field magnetic resonance; mapping the acquired K space data to a standard Cartesian coordinate system by adopting a grid interpolation method of a Gaussian kernel, and weighting each radial sampling point through a Gaussian kernel function to ensure that the data is smoothly distributed on a standard grid; image reconstruction is carried out based on image prior constraints by using the interpolated K space data, and the image prior constraints comprise a low-rank structure and sparse prior information of the image; parameters of low-rank and sparse components are optimized step by step by using a mixed kernel function, and constraint is carried out by combining a low-rank structure and sparse prior information of an image, so that the image reconstruction effect is improved. The invention also relates to a low-field magnetic resonance diffusion imaging reconstruction system. According to the invention, the image quality and the anti-noise performance can be obviously improved.

The invention provides a magnetic resonance diffusion weighted image denoising method and system, computer equipment and a medium, and belongs to the technical field of image processing. According to the method, a secondary moving grid division strategy is utilized, a standard deviation matrix is calculated according to texture features of a magnetic resonance diffusion weighted image, the image is divided into a flat area, a low-texture-richness area and a high-texture-richness area, and a corresponding adaptive mask matrix is generated. In the training stage, the image is masked by adopting a high mask rate; in the reasoning stage, a low mask rate is adopted, and the denoising stability is improved in combination with an average result of multiple mask reasoning. Through a denoising model constructed based on a blind spot network idea, a pixel value of a masked area is predicted by using pixels of an unmasked area, and model parameters are optimized by using an L2 loss function. By the adoption of the method, the definition of the image is improved, texture details are reserved, the method is not sensitive to the number of directions, and a stable and good denoising effect can be achieved when the number of directions is small.

The present application relates to a magnetic resonance imaging method, apparatus, computer device, and storage medium. The method includes: performing multiple excitations on a scanning object and acquiring multiple sets of diffusion-weighted echo signals, with each set of diffusion-weighted echo signals corresponding to one excitation; obtaining the phase information of each set of diffusion-weighted echo signals; correcting the phase information of each set of diffusion-weighted echo signals to obtain the corrected phase information of each set of diffusion-weighted echo signals; and performing image reconstruction based on each diffusion-weighted echo signal and the corrected phase information of each set of diffusion-weighted echo signals to obtain a magnetic resonance diffusion-weighted image. Using this method can improve the quality of the obtained magnetic resonance diffusion-weighted image.