Magnetic resonance fast imaging method and device based on convolutional neural network

Abstract

The invention discloses a magnetic resonance fast imaging method and device based on a convolutional neural network, and the method comprises the steps: collecting a magnetic resonance image, and carrying out the Fourier transformation of the magnetic resonance image, and obtaining k-space data; wherein the zero initialization length is the floating point type vector of the line number of the k space data to construct an image reconstruction network, sampling the k space data, performing inverse Fourier transform on the sampled k space data to obtain an image, and inputting the image into the image reconstruction network to obtain an output; calculating the L1 distance between the output of the image reconstruction network and the target image as a loss function; and obtaining a binary sampling vector according to the floating point type vector obtained by training, compiling a sampling sequence for the magnetic resonance instrument, and inputting the acquired magnetic resonance image into an image reconstruction network to obtain an output high-quality magnetic resonance image. In actual use, magnetic resonance images are collected according to a magnetic resonance sampling sequence, the images are input into a magnetic resonance image reconstruction network, and clear magnetic resonance images are obtained.

Description

technical field [0001] The invention relates to the technical fields of medical imaging and deep learning, in particular to a convolutional neural network-based fast magnetic resonance imaging method and device. Background technique [0002] In the field of fast magnetic resonance imaging, partial k-space data reconstruction can effectively speed up the imaging speed. The core idea is not to perform full sampling in the process of sampling k-space, but to selectively ignore the data of a part of the position, thereby reducing number of samples to reduce imaging time. Cartesian sampling is a commonly used undersampling mode, which samples an entire row or column of two-dimensional k-space data at a time. The sampled k-space data is inversely transformed by Fourier to obtain a magnetic resonance image. [0003] Due to the long scanning time of MRI, many advanced technologies such as cardiovascular imaging, functional magnetic resonance imaging, and magnetic resonance spectro...

AI Technical Summary

Problems solved by technology

[0003] Many advanced techniques such as cardiovascular imaging, functional magnetic resonance imaging, and magnetic resonance spectroscopy have not been widely used due to the long scanning time of magnetic resonance imaging.

Images