Deep Learning Artifact Correction for Synthetic MRI Contrast Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional MRI techniques require multiple scans to generate multiple image contrasts, leading to increased scan time and potentially lower quality synthetically generated MR contrast images due to artifacts, which hinder diagnostic value.
Innovation Solution
A deep learning-based method that utilizes a trained neural network to generate artifact-corrected reconstructed contrast images by inputting synthesized contrast images from a multi-delay multi-echo (MDME) scan and composite images derived from both MDME and contrast MRI sequences, with image intensity normalization before processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple separate MRI scans are conducted to generate multiple image contrasts, then diagnostic quality is improved, but scan time increases
Solution Approach 1:
The patent combines multiple MRI scan types (MDME quantitative scan and contrast-enhanced T1-weighted scan) into a single integrated scanning session, acquiring both structural and functional data simultaneously. This merging approach maintains comprehensive diagnostic quality while eliminating the time penalty of sequential scanning, directly resolving the contradiction between diagnostic quality and scan time.
Solution Approach 2:
The integrated scan protocol serves multiple diagnostic functions in one acquisition: generating T1 maps, T2 maps, proton density maps from the MDME sequence, and simultaneously acquiring contrast-enhanced T1-weighted images. This multi-functionality eliminates redundant scanning while preserving all necessary diagnostic information, addressing both the quality and time concerns.
2Loss of time
If synthetic MRI techniques are used to generate multiple contrasts from a single scan, then scan time is reduced, but image quality deteriorates due to artifacts
Solution Approach 1:
The patent introduces a deep learning-based artifact correction module as an intermediary processing step. This module takes the synthetically generated contrast images and removes artifacts through learned transformations, mediating between the efficient synthetic generation process and the high-quality diagnostic output requirement, thereby resolving the quality degradation issue.
Solution Approach 2:
The artifact correction network transforms the synthetic images by learning optimal parameter adjustments to eliminate artifacts. By changing the image parameters through learned transformations rather than direct synthesis, the system maintains diagnostic quality while preserving the time efficiency of single-scan approaches.
3Measurement precision
If deep learning-based artifact correction is applied to synthetic MRI images, then image quality is improved, but processing complexity increases
Solution Approach 1:
The artifact correction network is pre-trained on large datasets of paired synthetic and reference images before deployment. This preliminary training phase captures complex artifact patterns and correction strategies, allowing the network to perform rapid artifact removal during actual clinical use without requiring complex real-time processing, thus managing complexity effectively.
Data Source
AI summary
A computer-implemented method for generating an artifact corrected reconstructed contrast image from magnetic resonance imaging (MRI) data is provided. The method includes inputting into a trained deep neural network both a synthesized contrast image derived from multi-delay multi-echo (MDME) scan data or the MDME scan data acquired during a first scan of an object of interest utilizing a MDME sequence and a composite image, wherein the composite image is derived from both the MDME scan data and contrast scan data acquired during a second scan of the object of interest utilizing a contrast MRI sequence. The method also includes utilizing the trained deep neural network to generate the artifact corrected reconstructed contrast image based on both the synthesized contrast image or the MDME scan data and the composite image. The method further includes outputting from the trained deep neural network the artifact corrected reconstructed contrast image.