A large-view magnetic resonance scanning image reconstruction method and device based on deep learning

Abstract

The invention provides a large-visual-field magnetic resonance scanning image reconstruction method and device based on deep learning, and the method comprises the steps: obtaining a large-visual-field magnetic resonance scanning image which is an under-sampled scanning image; inputting the magnetic resonance scanning image into a pre-constructed deep neural network model; and reconstructing the magnetic resonance scanning image through the deep neural network model to obtain a high-resolution image corresponding to the under-sampled scanning image. By utilizing the technical scheme provided by the embodiment of the invention, the problem that the spatial resolution of the image is reduced due to the increase of the FOV after the selection of the scanning matrix in the conventional magnetic resonance scanning is solved, and a high-resolution image can be reconstructed in a relatively short time by utilizing a deep learning method on the premise of ensuring the large-view scanning imaging.

Description

technical field [0001] The present application belongs to the technical field of image processing, and in particular relates to a deep learning-based large-field magnetic resonance scanning image reconstruction method and device. Background technique [0002] At present, the severity of cerebral atherosclerosis is assessed clinically mainly by cerebral arterial angiography, by measuring the degree of stenosis of blood vessels. However, studies have found that during the development of atherosclerosis, the arterial wall will undergo positive remodeling, and the lesions leading to ischemic stroke are mainly located in the arterial vascular bed upstream of the brain tissue. If only vascular stenosis is tested, but Accurate detection cannot be performed without showing that there are lesions. [0003] Among the causes of ischemic stroke, intracranial artery disease accounts for 46.6%, and carotid artery disease accounts for about 30%. Plaque rupture triggers thrombosis, leading...

AI Technical Summary

Problems solved by technology

[0004]However, since the brain and carotid arteries need to be imaged at the same time, it is necessary to scan the two parts in one stop with a large field of view. In addition, the cerebral arteries are widely distributed and have many blood vessel branches. Scan coverage places higher demands on

The difficulty of head and neck integrated magnetic resonance imaging lies in the intracranial part. Early intracranial imaging was mostly two-dimensional imaging technology, through fast spin-echo sequences, with multi-layer crossing to improve coverage; while two-dimensional technology can only observe a section of the section For images, the layer thickness is generally too large and not isotropic, which cannot fully meet the needs of clinical applications

The current head and neck integrated imaging technology can obtain a maximum field of view of 250 mm, but considering factors such as the contrast of the scanned image, isotropic resolution, and scanning time, this field of view still cannot meet the clinical application requirements

In addition, after the scan matrix is ​​selected in the existing MRI scan, the larger the FOV will lead to the increase of the volume of the image voxel, the spatial resolution of the image will be reduced accordingly, so blindly increasing the FOV will lead to an increase in the spatial resolution of the MRI scan image. reduce

Images