43 results about "K space sampling" patented technology

In a magnetic resonance (MR) image reconstruction method and MR apparatus, for raw MR data are acquired with the propeller technique, and k-space sampling ensues in sub-data sets. The sample points of each sub-data set correspond to grid points of a Cartesian initial grid and the Cartesian initial grid of the sub-data sets can be brought into congruence by rotation: A Cartesian final grid is selected, and a calculation-based transfer of the data points of each sub-data set ensues to a respective new grid that exhibits the orientation of the respective sub-data set and the grid constants of the final grid, if the grid constants of the output grid differ from those of the final grid. A calculation-based transfer of the data points of each sub-data set, or the data points of a respective new grid (if obtained) ensures to the final grid by the application of a rotation module followed by transformation of the acquired data into the image domain.

The invention discloses a compressed sensing nuclear magnetic resonance imaging method based on a deep neural network. Through the method, a high-quality nuclear magnetic resonance image can be reconstructed with k-space low sampling data collected by a nuclear magnetic resonance imaging device. The method mainly comprises three steps: constructing an alternating direction multiplier deep neural network, training the parameters of the network, and applying the trained network to compressed sensing nuclear magnetic resonance imaging. A nuclear magnetic resonance image reconstructed with multiple pairs of sampling data at low sampling rate and corresponding full-sampling data is used as a training data set, and the model parameters of the alternating direction multiplier deep neural network are trained so that the output image of the deep neural network when sampling data at low sampling rate is used as input is as close as possible to an image reconstructed with full-sampling data. In application, k-space sampling data at given low sampling rate is input to a trained alternating direction multiplier deep neural network, and the output of the network is a reconstructed nuclear magnetic resonance image.

In a data acquisition and reconstruction method for magnetic resonance (MR) tomography, and a corresponding MR tomography apparatus, a blade-like sampling of k-space according to the PROPELLER method using a number of reception coils ensures with partial under-sampling of at least one blade of k-space such that the under-sampling ensues by regular omission of k-space lines in both boundary regions (with regard to the phase-encoding direction k_y) of a blade such that only data in each A-th line of said boundary regions are acquired; with no k-space lines being omitted in the central region (with regard to the k_y-direction) and thus at least one coil calibration line is obtained, selection of a suitable PPA method for completion of the blades and determination of the necessary coil calibration data necessary for the PPA reconstruction of a partial under-sampled blade from the central completely sampled region of said blade. PPA reconstruction via application of the selected PPA method selected in order to interpolate the non-measured or, respectively, omitted k-space lines of each blade, and execution of the PROPELLER reconstruction after the PPA reconstruction.

In a method and magnetic resonance apparatus for generation of a measurement sequence that can be executed with the hardware of the magnetic resonance apparatus, a measurement sequence is generated as a series of time slices of different time slice types, with the control signals for the magnetic resonance apparatus during each time slice being determined using parameters describing these time slices. After division of the measurement sequence into a number of time slices dependent on the sequence type and on a k-space sampling type, descriptive variables are associated with the time slices. Value ranges of the variables are limited and/or the variables are set in relation to one another using boundary conditions. A determination ensues of solution values of the variables with which a predetermined target parameter of the measurement sequence is optimized. A measurement sequence executable on the hardware is attained by association of the solution values with the corresponding parameters of the time slices.

The invention discloses a magnetic resonance frequency and phase position double-encoding sampling method and an image reconstruction method. The sampling method comprises the following steps of (a) setting an acquisition sequence and acquiring k-space data; (b) performing N-time over-sampling to the frequency encoding gradient direction according to the acquisition sequence during acquisition and simultaneously adding frequency encoding gradient and phase position encoding gradient to control data acquisition tracks; (c) filling the data acquired in the step (b) into a k-space through an analog to digital converter (ADC) to form the k-space with an inclination factor N, wherein data points of the k-space are distributed along the frequency encoding gradient direction and the phase position encoding gradient direction in an inclined N*m rows*n columns mode, and (N-1) blank data points are filled between every two adjacent actually acquired data points in the frequency encoding direction and the phase position encoding direction, the N is an integer larger than 1, and the m and the n are positive integers. The magnetic resonance frequency and phase position double-encoding sampling method enables phase positions and frequency to be encoded simultaneously so as to control the k-space sampling tracks, enables acquisition times to be less than the acquisition s times of a traditional acquisition s mode, shortens the total acquisition time and enables reconstructed images not to have image artifacts basically.

In a method and apparatus for accelerated spiral-coded imaging in magnetic resonance tomography using spiral-shaped k-space sampling, the underlying k-matrix is under-sampled such that an additional spiral is obtained by point mirroring of the measured values at the center of the k-matrix. This additional spiral forms a complete data set of the k-matrix together with the first spiral for imaging the output information.

The invention discloses a magnetic resonance dynamic imaging sampling method which includes the steps of acquiring multi-frame images and allocating k-space sampling point positions of different frameimages to form a position complementary or approximately complementary relationship, wherein after being combined, the sampling point positions of different frame images form a complete sampling or approximately complete sampling k-space, which is called a combined k-space. A variety of magnetic resonance reconstruction methods can be used to reconstruct images, such as a compressive sensing method, a method based on prior images or reference images, and a parallel reconstruction method (GRAPPA, SENSE, and variants thereof). The invention also discloses two methods for performing image reconstruction on the sampling method. Compared with a traditional under-sampling method, the technology can further reduce the sampling rate, reduce the number of samples, further reduce the scanning time,and improve the quality of reconstruction results.