71 results about "Parallel magnetic resonance imaging" patented technology

The present invention includes a method for parallel magnetic resonance imaging termed Superresolution Sensitivity Encoding (SURE-SENSE) and its application to functional and spectroscopic magnetic resonance imaging. SURE-SENSE acceleration is performed by acquiring only the central region of k-space instead of increasing the sampling distance over the complete k-space matrix and reconstruction is explicitly based on intra-voxel coil sensitivity variation. SURE-SENSE image reconstruction is formulated as a superresolution imaging problem where a collection of low resolution images acquired with multiple receiver coils are combined into a single image with higher spatial resolution using coil sensitivity maps acquired with high spatial resolution. The effective acceleration of conventional gradient encoding is given by the gain in spatial resolution Since SURE-SENSE is an ill-posed inverse problem, Tikhonov regularization is employed to control noise amplification. Unlike standard SENSE, SURE-SENSE allows acceleration along all encoding directions.

Methods and systems in a parallel magnetic resonance imaging (MRI) system utilize sensitivity-encoded MRI data acquired from multiple receiver coils together with spatially dependent receiver coil sensitivities to generate MRI images. The acquired MRI data forms a reduced MRI data set that is undersampled in at least a phase-encoding direction in a frequency domain. The acquired MRI data and auto-calibration signal data are used to determine reconstruction coefficients for each receiver coil using a weighted or a robust least squares method. The reconstruction coefficients vary spatially with respect to at least the spatial coordinate that is orthogonal to the undersampled, phase-encoding direction(s) (e.g., a frequency encoding direction). Values for unacquired MRI data are determined by linearly combining the reconstruction coefficients with the acquired MRI data within neighborhoods in the frequency domain that depend on imaging geometry, coil sensitivity characteristics, and the undersampling factor of the acquired MRI data. An MRI image is determined from the reconstructed unacquired data and the acquired MRI data.

In a first step of image formation by MRI or the like, a gridding convolution is performed for each coil using a Kaiser-Bessel kernel, so that each coil channel now contains data on grid points. In the second step, another 2D convolution is performed on each of the gridded coil data sets in either k-space or image space, and all the convolutions are summed up to produce the composite k-space data set for the target coil. In the third step, the composite data set is then transformed to form an un-aliased image for that coil. In the fourth step, the un-aliased images for all the coils are combined using square-root-of-sum-of-squares to form the final reconstructed image.

The invention relates to a Fourier parallel magnetic resonance imaging method based on a one-dimensional part of a deep convolutional network, and belongs to the technical field of magnetic resonance imaging. The method comprises the following steps: a sample set for training and a sample label set are created; an initial deep convolutional network is built; a training sample of the sample set is input into an initial deep convolutional network model to perform forward propagation, an output result of the forward propagation is compared with an expect result in the sample label set, and training is performed using a gradient descent algorithm until various layer parameters maximizing the consistency between the output result and the expect result are obtained; an optimal deep convolutional network model is established by utilizing the obtained various layer parameters; and a multi-coil under-sampling image obtained through online sampling is input into the optimal deep convolutional network model, forward propagation is performed on the optimal deep convolutional network model, and a rebuilt single-channel whole-sampling image is output. A noise of the rebuilt image can be removed well, a magnetic resonance image having a good visual effect is rebuilt, and the Fourier parallel magnetic resonance imaging method has high practical value.

The invention relates to a method of parallel imaging reconstruction in parallel magnetic resonance imaging reconstruction. Magnetic resonance data is acquired in parallel by an array of separate RF receiver coils. A reconstruction method based on Tikhonov regularization is presented to reduce the SNR loss due to geometric correlations in the spatial information from the array coil elements. In order to reduce the noise amplification of the reconstruction so-called “g-factor”, reference scans are utilized as a priori information of the final reconstructed image to provide regularized estimates for the reconstruction using the L-curve technique. According to the invention the method with the proposed L-curve approach was fully automatic and showed a significant reduction in average g-factors in the experimental_images.

The invention discloses a phase processing method for parallel magnetic resonance imaging. The phase processing method comprises the following steps of performing Fourier inverse transformation on K spacial data acquired by multi-channel coils in the parallel magnetic resonance imaging to obtain amplitudes and phases of all coil images; constructing a reference coil image, and estimating the spatial sensitivity distribution of all the coils in multiple channels; performing two-dimensional Fourier transformation on the spatial sensitivity distribution of the reference coil image, and intercepting an intermediate matrix as a convolution kernel; constructing a K spacial data convolution model, and solving a joint weight W of the coils; obtaining a K spacial value of a virtual coil and performing Fourier inverse transformation to obtain a virtual coil image; unwrapping a phase and removing the phase of the background of the virtual coil image; extracting the phase of a region of interest by using a mask image. According to the phase processing method disclosed by the invention, phase information of the image is acquired by combining K space with coil data, and the phenomenon that a phase information acquisition algorithm based on an image domain is influenced by noise and aliasing artifact in the reconstruction of the parallel magnetic resonance imaging under the condition of accelerated sampling is avoided.

The invention discloses a parallel magnetic resonance imaging GRAPPA (generalized autocalibrating partially parallel acquisitions) method based on machine learning. The method comprises the steps of: acquiring K spatial data set from a to-be-imaged object, creating mapping relation between an undersampled point and a neighbor point by virtue of regression analysis in the machine learning, predicting the undersampled point, and filling up the undersampled K space, performing Fourier inverse transformation to K spatial data of each coil to obtain the image of each coil, and solving quadratic sum of multiple images to obtain the last reconstructed result. Based on the method, the mapping relation between the undersampled point and the neighbor point is estimated by virtue of the regression analysis in the machine learning, and the linear mapping relation in the original algorithm is replaced, and the undersampled space is filled up, at last the more accurate reconstructed result can be obtained, so that the artifact of the magnetic resonance reconstructed image can be reduced.

A parallel magnetic resonance imaging (MRI) apparatus configurable to image a physical entity comprises: a main magnetic flux source for providing a uniform fixed magnetic field, Balpha; an RF array system comprising a plurality of RF coils and receivers, said RF system configured for: generating rotating RF excitation magnetic fields B1; and receiving RF signals due to precessing nuclear magnetization on multiple spatially distinct radio frequency coils and associated receiver channels, said RF system being configured to operate in accordance with a B1 sensitivity encoding technique; a control processor for controlling imaging functionality, collecting image data and effecting data processing of the captured image data the control processor being configured with post processing capability for the B1 sensitivity encoding technique; an image display means for displaying processed image data as resultant images; and an auxiliary magnetic field means capable of producing at least one auxiliary uniform Bo step magnetic field imaging region within the main B0 magnetic field; wherein: the auxiliary magnetic field, means is configured to operate in combination with the RF coil system and the B1 sensitivity encoding technique, the imaging apparatus thereby providing faster image acquisition than that attributed to the speed up factor provided solely by the B1 sensitivity encoding technique. The invention also includes a method of imaging using this apparatus. Furthermore, the invention also includes a method and apparatus for three-dimensional MR imaging using a 1D Multiple Acquisition Micro Bo array coupled with a 2D Multiple Acquisition Micro Bo array.

A variable density non-Cartesian parallel-imaging method for reconstructing a magnetic resonance (MR) image is provided. In embodiments of the invention, an MR data set is obtained by sampling first and second sampling regions, wherein a first region is sampled with a first sampling density that is higher than a second sampling density of a second region. MR images corrupted by aliasing artifacts are reconstructed from the data obtained with each one of the coil-elements of a coil array. These images can be combined into one, de-aliased image using a modified version of Cartesian SENSE. The modification allows all the available k-space lines to be used in the processing, despite the fact that different k-space regions have different sampling densities (i.e. non-Cartesian sampling). Using all available lines is advantageous in terms of signal-to-noise ratio. Advantages of embodiments of the invention over previous methods also able to deal with non-Cartesian sampling schemes may include one or more of simplicity, ease of implementation, not having to fit sensitivities to target functions as part of the reconstruction, fast processing speed and/or the avoidance of possible errors resulting from solving large systems of equations.

The application relates to through-time non-cartesian GRAPPA calibration. Example systems and methods control a parallel magnetic resonance imaging (pMRI) apparatus to acquire radial calibration data sets throughout time. Example systems and methods also control the pMRI apparatus to acquire an under-sampled radial data set from the object to be imaged. Example systems and methods then control the pMRI apparatus to reconstruct an image of the object to be imaged from the under-sampled radial data set. The reconstruction depends, at least in part, on a through-time radial GRAPPA calibration where a value for a point missing from k-space in the under-sampled radial data set is computed using a GRAPPA weight set calibrated and applied for the missing point. The GRAPPA weight set is computed from data in the radial calibration data sets.

The invention belongs to the field of magnetic resonance imaging, discloses an improved reconstruction method for parallel magnetic resonance images and particularly relates to a reconstruction method suitable for K spatial data in the parallel magnetic resonance imaging process. According to the method, when a weight coefficient is fitted, estimation is carried out with a K spatial data area and a self-calibration data area near undersampled data instead of adjacent points and a self-calibration line, and thus the nonlinear relation of K spatial data can be better fitted; noise in the sampled signals can be suppressed in the singular value decomposition and truncation processing process, and it is ensured that the signal-to-noise ratio of images is high. The method is a GRAPPA improving method based on matrix generalized inverse and singular value decomposition, the noise of the K spatial data can be suppressed, the fitting precision can be improved, and meanwhile the imaging quality can still be ensured when an acceleration factor is large.

A parallel magnetic resonance imaging device comprises a plurality of imaging channels, a collection module, an initialization module, an operation module and a reconstruction module, wherein the collection module is used for collecting an undersampling matrix di (i is the serial number of each imaging channel and i > 0) from the plurality of imaging channels in an undersampling method and according to undersampling factors, the initialization module is used for acquiring an initialization image matrix rho and an initialization sensitivity matrix si (i is the serial number of each imaging channel and i > 0) of the imaging channels, the operation module is used for optimization iteration solving of a constraint function by means of the conjugate gradient algorithm according to the undersampling matrix di, the initialization image matrix rho and the initialization sensitivity matrix si to obtain a reconstructed image matrix rho and a sensitivity matrix si, and the reconstruction module is used for reconstructing an image according to the reconstructed image matrix rho and the sensitivity matrix si. When the parallel magnetic resonance imaging device is used for image reconstruction, imaging speed is effectively increased, and signal to noise ratio loss is small.

A method for reconstructing a full k-space dataset using parallel magnetic resonance (MR) imaging technique is provided. The method includes acquiring, by a plurality of receiver coils, a set of first under-sampled k-space data, receiving a set of second partial or fully-sampled k-space data, respectively performing k-space interpolation of the set of the first under-sampled k-space data respectively acquired by each of the plurality of receiver coils, recovering respectively missing k-space lines of each of the set of first under-sampled k-space data using corresponding second partial or fully-sampled k-space data and corresponding first under-sampled k-space data, forming a plurality of full k-space datasets by respectively combining each of the set of first under-sampled k-space data and corresponding recovered missing k-space lines for each of the plurality of receiver coils, obtaining a plurality of fully-sampled images from the plurality of full k-space datasets, and combining images into a final image.

The invention discloses a parallel magnetic resonance imaging high quality reconstruction method based on self-consistency and containing combined total variation; the method is based on a SPIRiT frame, and provides the high quality reconstruction method aiming at parallel imaging reconstruction problems containing JTV and JL1 composite regular terms; firstly a constraint reconstruction problem is converted into an unconstrained reconstruction problem; then a data fidelity term and a self-correction term are simplified; then the operator split technology converts the simplified reconstruction problem into a gradient calculating problem and a denoising problem containing the JTV and JL1 composite regular terms, and the composite regular terms denoising problem is solved through a whole novel design algorithm based on Split Bregman technology; finally FISTA can accelerate the speed. The reconstruction performances of the novel algorithm and other normal algorithms are compared; test simulation shows that the convergence speed of the novel algorithm is close to a POCS algorithm, yet reconstruction image SNR can be largely improved.

The invention provides a parallel magnetic resonance imaging method and device based on adaptive joint sparse coding and a computer readable medium. According to the method, a l2-lF-l2, 1 minimum objective function is solved, wherein the l2 norm is a data fitting item, the lF norm represents a sparse representation error, and the l2, 1 mixed norm represents a joint sparse constraint between channels; then updating a sparse matrix, a dictionary and K-space data by adopting a divide-and-rule method and a corresponding algorithm; and finally solving a reconstructed image according to the sum of root mean squares of all channels. The method provided by the invention develops the joint sparsity of the channels by using the l2, 1 norm, calibration can be removed while the information sparsity isdeveloped, and the method has high robustness.

In a method and magnetic resonance apparatus for acquiring a sensitivity map for at least one local coil in a magnetic resonance scanner, the extent of k-space to be sampled is divided into a first part located around the center of k-space, and a second part. First and the second magnetic resonance data sets are acquired with undersampling in at least one phase-coding direction in the second part, and are acquired globally in the first part. An accelerated parallel magnetic resonance imaging reconstruction technique is executed for the reconstruction of magnetic resonance data that are missing in the magnetic resonance raw data sets due to the undersampling, to produce a global data set defined by combining the first and the second magnetic resonance global data sets. Supplemented first and second magnetic resonance data sets are acquired by adding the reconstructed magnetic resonance data in the regions not covered in the undersampling. The sensitivity maps are acquired from the magnetic resonance data sets that have been supplemented in this way.

The invention relates to a self-consistent parallel magnetic resonance imaging reconstruction method combining total variation Lp pseudo norms based on feature vectors, and belongs to the technical field of medical magnetic resonance imaging. The invention provides an ESPIRiT parallel imaging reconstruction algorithm containing a joint total variation Lp pseudo norm regularization term based on aniteration self-consistency parallel imaging reconstruction ESPIRiT framework of a feature vector. The method comprises the following steps: firstly, decomposing a reconstruction problem into a gradient calculation problem and a denoising problem containing sparse regular terms by adopting an OS technology; secondly, denoising is carried out by applying an MM algorithm; and finally, FISTA is usedfor acceleration. Experimental results show that compared with a traditional reconstruction algorithm using L1 regular terms, the proposed new algorithm using LpJTV regular terms based on the ESPIRiTreconstruction model can more effectively improve the reconstruction quality of the image.

The invention discloses a self consistency based parallel magnetic resonance imaging quick reconstructing method. Based on an SPIRiT frame, aiming at a JTV and JL1 composite regular term contained parallel imaging reconstruction problem, the invention proposes a quick reconstructing method employing Descartes sampling. In the prior art, scanning time can be reduced distinctively by adopting the magnetic resonance imaging technology; however, the reconstructing time of a prior reconstructing algorithm is long. The method provided by the invention utilizing a split Bregman technology for dissolving a prior problem into a plurality of sub problems that are easy to solve. Experiment simulation results show that the new algorithm is greatly improved in convergence speed under the premise of ensuring reconstruction quality compared with a prior NLCG algorithm.