Simultaneous multi-slice magnetic resonance parameter quantification imaging method

CN117017263BActive Publication Date: 2026-07-07XIAMEN UNIV
View PDF 4 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
XIAMEN UNIV
Filing Date
2023-08-17
Publication Date
2026-07-07

AI Technical Summary

Technical Problem

Traditional quantitative magnetic resonance imaging methods require multiple scans to obtain images with different weights, resulting in long data acquisition times and limiting their clinical application. Furthermore, existing single-scan methods can only image one layer, and the total acquisition time increases with the number of layers when scanning layer by layer.

Method used

The design incorporates a multi-layer magnetic resonance parameter quantitative imaging pulse sequence, combined with a deep neural network. By acquiring k-space data in a single scan and performing preprocessing, multi-layer magnetic resonance parameter quantitative images are reconstructed. The design employs a multi-layer signal excitation module, a shift gradient module, a coil sensitive map pre-scanning module, and a deep neural network reconstruction algorithm.

Benefits of technology

It enables quantitative imaging of multi-slice magnetic resonance parameters in a single scan, shortens the acquisition time, is suitable for low and high acceleration magnification, and is applicable to multi-slice quantitative imaging of 4 layers and above, thus improving imaging efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117017263B_ABST
    Figure CN117017263B_ABST
Patent Text Reader

Abstract

The application relates to a simultaneous multi-layer magnetic resonance parameter quantitative imaging method, which relates to magnetic resonance imaging. The method comprises the following steps: S1, designing a simultaneous multi-layer magnetic resonance parameter quantitative imaging pulse sequence; S2, generating a training sample of a deep neural network; S3, training the deep neural network by using the training sample to obtain a trained deep neural network; S4, collecting k-space data of an actual imaging object by using the simultaneous multi-layer magnetic resonance parameter quantitative imaging pulse sequence and carrying out pretreatment to obtain an aliasing image of the actual imaging object; and S5, reconstructing the aliasing image of the actual imaging object by using the trained deep neural network to obtain a multi-layer magnetic resonance parameter quantitative image. The method can realize multi-layer magnetic resonance parameter quantitative imaging in one scanning, and can not only be used for a lower acceleration multiple, but also be suitable for a high acceleration multiple, so that the acquisition time of quantitative imaging is shortened.
Need to check novelty before this filing date? Find Prior Art