Dynamic magnetic resonance image reconstruction method and device based on sparse and structured low rank

CN117665678BActive Publication Date: 2026-06-05INST OF ADVANCED TECH UNIV OF SCI & TECH OF CHINA +1

Patent Information

Authority / Receiving Office: CN · China
Patent Type: Patents(China)
Current Assignee / Owner: INST OF ADVANCED TECH UNIV OF SCI & TECH OF CHINA
Filing Date: 2023-12-01
Publication Date: 2026-06-05

AI Technical Summary

Technical Problem

The long scanning time of dynamic MRI technology affects the reconstruction quality and limits the spatiotemporal resolution. Reducing the sampling time and quickly reconstructing high-quality images is a challenge.

Method used

An image reconstruction method based on sparse and structured low-rank is adopted. By constructing a structured low-rank matrix and compressed sensing algorithm, combined with the alternating direction multiplier method, the image reconstruction algorithm is decomposed. Fourier transform and random singular value decomposition are used to optimize the reconstruction process.

Benefits of technology

It improves the reconstruction speed and quality of dynamic MRI images, makes full use of the interrelationships between magnetic resonance signals, and improves the temporal and spatial resolution of image reconstruction.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure CN117665678B_ABST

Patent Text Reader

Abstract

The present disclosure provides a dynamic magnetic resonance image reconstruction method and device based on sparse and structured low rank, and an electronic device thereof. The method comprises: grouping multi-channel sampling data obtained from a database according to a preset image frame number to obtain grouped multi-channel sampling data, the multi-channel sampling data being stored in the database after being radially sampled by a multi-channel coil of a dynamic magnetic resonance scanner when scanning a target object within a preset time period; constructing a structured low rank matrix based on the grouped multi-channel sampling data, the structured low rank matrix being used to represent a low rank characteristic of multi-channel Fourier space data; inputting the grouped multi-channel sampling data after Fourier transform into an image reconstruction algorithm to output reconstructed image domain data corresponding to each group of multi-channel sampling data, the image reconstruction algorithm being constructed based on a low rank constraint on the structured low rank matrix and a compressed sensing algorithm; and reconstructing images of the preset image frame number according to the reconstructed image domain data.

Need to check novelty before this filing date? Find Prior Art