A Dixon water/fat separation technique, in particular in combination with a single-point acquisition scheme, avoids swaps of water and fat signals in the reconstructed MR images due to imperfections of the main magnetic field B.sub.0. Echo signals are generated and acquired in a pre-scan by subjecting an object (10) to a two or more point imaging sequence. A fat fraction map is derived from the echo signals of the pre-scan. Echo signals are generated and acquired in a clinical scan by subjecting the object (10) to a single-point imaging sequence. A field map estimate is derived from the fat fraction map and from the echo signals of the clinical scan. An MR image is reconstructed from the echo signals of the clinical scan. Signal contributions from fat and water are separated on the basis of the field map estimate.

A method for magnetic resonance imaging (MRI) may include cause, based on a pulse sequence, a magnetic resonance (MR) scanner to perform a scan on an object. The pulse sequence may include a steady-state sequence and an acquisition sequence that is different from the steady-state sequence. The steady-state sequence may correspond to a steady-state phase of the scan in which no MR data is acquired. The acquisition sequence may correspond to an acquisition phase of the scan in which MR data of the object is acquired. The method may also include generating one or more images of the object based on the MR data.

The invention provides for a magnetic resonance imaging system (100). Machine executable instructions (140) cause a processor controlling the magnetic resonance imaging system to control (200) the magnetic resonance imaging system with the pulse sequence commands to acquire two point Dixon magnetic resonance data and single point Dixon magnetic resonance data; calculate (202) a first resolution magnetic field inhomogeneity map (148) using the two point Dixon magnetic resonance data; calculate (204) a second resolution magnetic field inhomogeneity map (154) by interpolating the first resolution magnetic inhomogeneity map to the second resolution; and calculate (206) a second resolution water image (156) and a second resolution fat image (158) using the single point Dixon magnetic resonance imaging data and the second resolution magnetic field inhomogeneity map. The first resolution is lower than the second resolution

The invention relates to a method of Dixon-type MR imaging. It is an object of the invention to provide a method that enables efficient and reliable water/fat separation. The method of the invention comprises the following steps: subjecting an object (10) to an imaging sequence, which comprises at least one excitation RF pulse and switched magnetic field gradients, wherein two echo signals, a first echo signal and a second echo signal, are generated at different echo times (TE1, TE2), acquiring the echo signals from the object (10), reconstructing a water image and/or a fat image from the echo signals, wherein contributions from water and fat to the echo signals are separated using a two-point Dixon technique in a first region of k-space and a single-point Dixon technique in a second region of k-space, wherein the first region is different from the second region. In other words, the invention proposes an adaptive switching between a two-point Dixon technique for water/separation, applied to both the first and second echo signals, and a single-point Dixon technique applied to one of the two echo signals, i.e. the first echo signal data or the second echo signal data, depending on the position in k-space. Moreover, the invention relates to a MR device (1) and to a computer program to be run on a MR device (1).

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.

A computer implemented method for measuring T.sub.1 in an anatomical region of interest during a dynamic procedure includes acquiring a reference MR image of the anatomical region of interest using a first flip angle. A first set of dynamic MR images of the anatomical region of interest are acquired using a second flip angle. The reference MR image and the first set are used to calculate a reference T.sub.1 value for tissue in the anatomical region of interest. During an intervention where the T.sub.1 value may change, a second set of dynamic MR images of the anatomical region of interest is acquired using the second flip angle. The reference MR image and the second set are used to calculate an estimated T.sub.1 value. The reference T.sub.1 value, the estimated T.sub.1 value, and the first and second flip angles may then be used to correct the estimated T.sub.1 value.

In a method and computer for creating a pulse sequence for controlling a magnetic resonance (MR) tomography system to generate image data, raw MR data are acquired by exciting different transverse magnetizations in a number of sub-volumes of the subject, with a sequence of pulse iterations being executed that each prepare, excite and read out sub-volumes. The pulse iterations are designed so that a readout occurs when the pulse sequence is applied between a preparation of two spatially directly adjacent sub-volumes.

The invention provides for a medical imaging system (100, 300). The medical imaging system (100, 300) comprises a processor (104). Execution of machine executable instructions(120) causes the processor (104) to: receive magnetic resonance data, wherein the magnetic resonance datacomprises B0 field data (122) of a reference scan for a plurality of voxels and water saturation data (124) of a WASSR scan for a subset of voxels of the plurality of voxels, the water saturation data (124) comprising data of a limited number of sample points; determine a local absolute water saturation frequency (130) for each voxel of the subset using the water saturation data (124) of the WASSR scan; and re-construct a field map (132) comprising a local absolute water saturation frequency for each voxel of the plurality of voxels, wherein the reconstruction comprises determining relative frequency differences between the voxels using the B0 field data (122) of the reference scan and adding a frequency offset to the relative frequency differences based on the determined local absolute water saturation frequencies (130) of the subset.

A method is used to carry out a magnetic resonance measurement with at least one echo train with n spin echoes and prospective movement correction. Movement correction data for each echo train is updated at the start of the echo train and is then updated again at most partially for the spin echoes.

A system and method is provided for acquisition of magnetic resonance fingerprinting (MRF) data that includes determining a non-locally sequential sampling pattern for a Cartesian grid of k-space, performing a series of sequence blocks using acquisition parameters that vary between sequence blocks to acquire MRF data from a subject using the Cartesian grid of k-space and the determined non-locally sequential sampling pattern, assembling the MRF data into a series of signal evolutions, comparing the series of signal evolutions to a dictionary of known signal evolutions to determine tissue properties of the subject, and generating a report indicating the tissue properties of the subject.