Rapid imaging method, device and equipment for nuclear magnetic resonance image

Abstract

A rapid imaging method of the nuclear magnetic resonance image comprises the steps of obtaining a full-sampling MR image, obtaining an under-sampling MR image according to the full-sampling MR image,and establishing a first training data set of mapping from the under-sampling MR image to the full-sampling MR image; Performing unified dimension processing and smooth processing on the first training data set to obtain a second training data set; A preset convolutional neural network is trained through the second training data set, the convolutional neural network is a U-shaped convolutional neural network with introduced residual terms, and trained VGG network parameters are loaded for transfer learning to initialize the network parameters; And reconstructing an MR image according to the trained convolutional neural network. The accuracy of network classification can be effectively improved, the initialization speed of network parameters is reduced, and the training efficiency of the network is improved.

Description

technical field [0001] The present application belongs to the field of image processing, and in particular relates to a fast imaging method, device and equipment for nuclear magnetic resonance images. Background technique [0002] Magnetic resonance imaging (MRI) has been applied to the imaging diagnosis of various systems of the whole body. It can display fat, whole body organs, muscles, fast-flowing blood, bones and air, etc., and can also clearly display the internal structure of organs. For cardiovascular diseases, it can not only observe the anatomical changes of chambers, great blood vessels and valves, but also can be used for ventricle analysis for qualitative and semi-quantitative diagnosis. It can make multiple slices with high spatial resolution to display the heart and lesions The overall picture and its relationship with the surrounding structures are superior to other X-ray imaging, two-dimensional ultrasound, nuclide and CT examinations. Doctors can very well...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the embodiment of the present application provides a fast imaging method, device and equipment for nuclear magnetic resonance images to solve the problem that in the prior art, it takes a long time to generate an MR image through a neural network, and the prediction accuracy of the generated MR image is low. The problem

Images