MRI Data Processing Using Trained Model for Image Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance imaging (MRI) techniques face a trade-off between imaging time and image quality, with conventional spin echo methods providing high image quality but long imaging times, and high-speed spin echo methods offering shorter times but lower quality due to the filling of k-space with data from multiple echo times.
Innovation Solution
A data processing apparatus and method that utilizes a trained model, specifically a deep neural network, to process MRI data from high-speed spin echo methods to generate output data with high image quality similar to conventional spin echo methods, by selecting and processing data segments corresponding to specific echo times, thereby improving image quality while maintaining shorter imaging times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional spin echo method is used, then image quality is high, but imaging time is long
Solution Approach 1:
The patent segments the k-space data into multiple regions (central region and peripheral regions) and processes each region differently. The central region data is used to generate an initial image, while peripheral region data is used to create correction images that are superimposed to enhance overall image quality. This segmentation allows the system to achieve high image quality comparable to conventional spin echo while maintaining the shorter imaging time of high-speed methods.
2Loss of time
If high-speed spin echo method is used, then imaging time is short, but image quality is low
Solution Approach 1:
The patent introduces correction images as an intermediary element. These correction images are generated from peripheral k-space data and are superimposed onto the initial image to compensate for the degraded quality caused by using high-speed spin echo acquisition. This intermediary approach allows the system to maintain short imaging times while recovering image quality through the corrective action of the superimposed images.
3Productivity
If high-speed spin echo method is used, then productivity is improved, but measurement precision deteriorates
Solution Approach 1:
The patent merges multiple image components (initial image from central k-space data and correction images from peripheral k-space data) to produce a final high-quality image. This merging process combines the advantages of fast acquisition with the quality requirements, allowing the system to achieve both high productivity through rapid data acquisition and high measurement precision through the combined image reconstruction approach.
Data Source
AI summary
According to one embodiment, a data processing apparatus includes processing circuitry. The processing circuitry acquires input data relating to a processing target including a plurality of data segments corresponding respectively to a plurality of imaging contrasts determined by a first pulse sequence. The processing circuitry generates output data relating to the processing target by applying a trained model to input data relating to the processing target. The processing circuitry outputs output data relating to the processing target.