Magnetic resonance image reconstruction method and device, equipment and medium

Abstract

The embodiment of the invention discloses a magnetic resonance image reconstruction method and device, equipment and a medium. The method comprises the following steps: acquiring undersampled magneticresonance data; and inputting the magnetic resonance data into an image reconstruction model based on an alternating direction multiplier algorithm to obtain a reconstructed target magnetic resonanceimage, the image reconstruction model being a pre-trained result of a mathematical model in which a data fidelity term and a regular term are both uncertain terms. According to the embodiment of theinvention, the method solves a problem that a data fidelity term in a model suitable for an ADMM-net method is represented by employing a 2-norm between a reconstructed k space and a sampling point ina magnetic resonance image reconstruction process, is not a most effective data consistency guarantee method, and is not suitable for all image reconstruction conditions. The image quality after image reconstruction based on the ADMM algorithm can be improved.

Description

technical field [0001] Embodiments of the present invention relate to medical imaging technology, and in particular to a magnetic resonance image reconstruction method, device, equipment and medium. Background technique [0002] Magnetic resonance uses static magnetic field and radio frequency magnetic field to image human tissue. It not only provides rich tissue contrast, but also has no side effects on human body, so it has become a powerful tool for medical clinical diagnosis. [0003] In order to improve the speed and imaging quality of magnetic resonance imaging, deep learning methods are often used for image reconstruction, such as using neural networks to learn the optimal parameters required for reconstruction from a large number of training data or directly learning from under-acquisition data to full-acquisition images The mapping relationship between them can achieve better imaging quality and higher acceleration than traditional parallel imaging or compressed sen...

AI Technical Summary

Problems solved by technology

[0005] However, the data fidelity item in the model applicable to the ADMM-net method is characterized by the 2-norm between the reconstructed k-space and the sampling point. This least square constraint is based on the premise of linear unbiased estimation. It is not the most effective data consistency guarantee method, and it is not suitable for all image reconstruction situations, resulting in the imaging quality of the reconstructed image to be improved

Images