A Deep Learning-Based Supercomplex Magnetic Resonance Spectrum Reconstruction Method

Abstract

A deep learning-based super-complex magnetic resonance spectrum reconstruction method involves magnetic resonance spectroscopy. The fully-sampled time-domain signal of the hyper-complex MR spectrum is generated according to the MR spectral signal index model, the fully-sampled time-domain signal is under-sampled according to the under-sampling template, and the positions of the data points that have not been collected are set to zero to obtain the hyper-complex MR. The under-sampling time-domain signal of the spectrum, the under-sampling time-domain signal and the corresponding full-sampling time-domain signal are replaced by complex time-domain signals, and the respective under-sampling and full-sampling magnetic resonance spectra in the frequency domain are obtained by Fourier transform. using the training set for hypercomplex spectral reconstruction; construct a deep learning network for hypercomplex MR spectral reconstruction, use the obtained training set to train the deep learning network, and obtain the trained network parameters for hypercomplex MR spectral reconstruction; The trained network reconstructs the undersampled time-domain signal of the hypercomplex MR spectrum to obtain a complete hypercomplex MR spectrum.

Description

technical field [0001] The invention relates to magnetic resonance spectroscopy, in particular to a deep learning-based super-complex magnetic resonance spectroscopy reconstruction method. Background technique [0002] Magnetic Resonance Spectroscopy (MRS) can unambiguously elucidate molecular structures and has been widely used in medicine, chemistry and biology. In undersampled MR spectral reconstruction, how to reduce the sampling time while ensuring the quality of the reconstructed spectrum is an important technical issue. [0003] Traditional magnetic resonance reconstruction methods mainly use the mathematical properties of magnetic resonance time or frequency signals to reconstruct the frequency spectrum. Qu Xiaobo et al. (Qu X, Mayzel M, Cai J, Chen Z, Orekhov V. Accelerated NMRspectroscopy with low-Rank reconstruction[J]. Angewandte Chemie International Edition, 2015, 54(3):852-854.) proposed a The magnetic resonance spectrum reconstruction method based on low-ran...

AI Technical Summary

Problems solved by technology

However, the singular value decomposition in the iterative process of this kind of low-rank Hankel matrix reconstruction method consumes a lot of time, thus resulting in a long spectral reconstruction time

Guo Di et al. (Guo D, Lu H, Qu X.A fast low rank Hankelmatrix factorization reconstruction method for non-uniformly sampled magneticresonance spectroscopy[J].IEEE Access,2017,5:16033-16039.) successfully decomposed the low rank matrix And introduce parallel computing to avoid singular value decomposition with high time complexity, but it still takes a certain amount of time

Images