Magnetic resonance rapid parameter imaging method and system

Abstract

The invention provides a magnetic resonance rapid parameter imaging method comprising the following steps: original K-space data inputted by a magnetic resonance scanner are received; a deep learning model is established according to a nerve network; training data are acquired, and Fourier inverse transformation is performed on the training data so that a training matrix is obtained; proofreading data are acquired, and Fourier inverse transformation and curve fitting are performed on the proofreading data so that a proofreading image is obtained, and a proofreading matrix is obtained according to the proofreading image; the deep learning model is trained according to the training matrix and the proofreading matrix so that the trained deep learning model is obtained; and a new training matrix is acquired, and a magnetic resonance parameter image is reconstructed according to the new training matrix and the trained deep learning model. Quality of the reconstructed image can be effectively enhanced by using the method. Besides, the invention also provides a magnetic resonance rapid parameter imaging system.

Description

technical field [0001] The invention relates to the field of magnetic resonance imaging, in particular to a magnetic resonance fast parameter imaging method and system. Background technique [0002] Magnetic resonance parametric imaging technology usually changes the parameters of the same imaging sequence, and acquires a series of parameter-weighted images related to the imaging parameters. For each pixel in the image, according to a certain formula, the corresponding parameters are obtained by fitting. However, parametric imaging requires multiple scans, and the imaging time is several to tens of times that of the structural image, and the speed is very slow. [0003] In order to meet the requirements of parametric imaging speed, K-space undersampling schemes such as parallel imaging technology and partial Fourier technology have been applied to parametric imaging, and the purpose of reducing scanning time is achieved by reducing the number of scanning lines in the phase ...

AI Technical Summary

Problems solved by technology

The commonly used MRI fast parametric imaging technology first reconstructs the parameter-weighted image from the downsampling data, and then uses the parameter-weighted image to fit the parameter values. Therefore, the error introduced when reconstructing the parameter-weighted image will be enlarged in the process of fitting the parameter value. Form error propagation, resulting in errors that degrade the quality of the reconstructed image

Images