Magnetic resonance parallel imaging method of multi-support vector model

Abstract

The invention discloses a magnetic resonance parallel imaging method of a multi-support vector model, which belongs to the field of magnetic resonance parallel imaging. The method comprises completely sampling an intermediate region in a K space, dividing into a training set and a testing set, performing accelerated sampling on the other regions to obtain samples as a prediction set, and performing normalization processing on the data in the sets; dividing the training set into a plurality of training subsets, and selecting different parameters to train each training subset by using a support vector to obtain different combined weighting function models; testing the combined weighting functions on the testing set, and selecting a plurality of optimal submodels; and predicting the prediction set by using the optimal submodels, taking an average value as a value of uncollected points, performing reverse normalization processing, and converting K space data into an image. The parallel imaging method has good generalization and small overall reconstruction error by using weighting functions fitted by the support vector.

Description

technical field [0001] The invention belongs to the field of magnetic resonance parallel imaging, in particular to a magnetic resonance parallel imaging method of multiple support vector machine models. Background technique [0002] Magnetic resonance imaging (MRI) has become one of the important means of clinical medical imaging due to its advantages of no nuclear radiation, high resolution, and multi-directional and multi-parameter imaging. However, limited by the Fourier encoding method and the Nyquist sampling theorem, the speed of MRI is slow, which not only brings some discomfort to patients, but also easily produces motion artifacts. At the same time, the long scanning time limits MRI imaging of moving objects, such as infants, blood flow, heart, etc. After decades of development, relying on improving hardware performance to accelerate acquisition has reached the limit of the human body. [0003] Parallel imaging technology uses multiple coils to acquire signals at ...

AI Technical Summary

Problems solved by technology

How to select the best adjacent subset and the size of this set has always been the difficulty of the GRAPPA algorithm. Some methods use iterative or cross-validation methods to select the adjacent subset with the smallest error, but the calculation time is too long to satisfy real-time requirements

At the same time, solving by the least squares method can only minimize the matching error in the calibration area, without considering the complexity of the joint weight function, which is prone to over-matching, resulting in an increase in the prediction error of unacquired points

Images