Motion Field Generation for PET-MRI Image Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion artifact correction techniques for PET-MRI systems either limit image reconstruction to specific respiratory phases, resulting in low signal-to-noise ratio, or rely on special MRI sequences that cannot efficiently correct image data acquired outside these sequences, especially when the object's motion is unstable.
Innovation Solution
A method for generating motion fields based on MR image data and a motion curve associated with physiological motion, allowing for the correction of PET and MR images across the entire scanning time period, even when special MRI sequences are not used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If motion waveform guided reconstruction technique is used, then motion artifact correction is achieved for specific respiratory phase, but signal-to-noise ratio of corrected image is low
Solution Approach 1:
The patent combines multiple respiratory phase image data into a single corrected image through merging. Instead of reconstructing images for each respiratory phase separately (which results in low SNR), the system acquires image data across multiple respiratory phases and merges them using calculated motion compensation factors, thereby improving the signal-to-noise ratio while maintaining motion artifact correction accuracy.
Solution Approach 2:
The patent creates a universal motion compensation method that works across all respiratory phases rather than being limited to a specific phase. By calculating motion compensation factors that apply universally to image data from different respiratory phases, the system achieves both motion artifact correction and high SNR in the final corrected image.
2Reliability
If special MRI sequences are used for motion field generation, then motion artifact correction is achieved between different respiratory phases, but other image data acquired when special sequences are not used cannot be efficiently corrected
Solution Approach 1:
The patent develops a universal motion field generation method that does not depend on special MRI sequences. The system calculates motion fields using standard imaging sequences combined with respiratory waveform data, making the motion artifact correction technique applicable to all image data acquired during the scan regardless of whether special sequences were used.
Solution Approach 2:
The patent introduces respiratory waveform data as an intermediary to bridge the gap between image acquisition and motion correction. Instead of relying on special MRI sequences to directly provide motion fields, the system uses readily available respiratory waveform data combined with standard imaging sequences to generate the necessary motion compensation information.
3Reliability
If motion field from special MRI sequences is used to correct other image data, then correction is attempted, but correction efficiency is low due to unstable motion
Solution Approach 1:
The patent implements a feedback mechanism where motion compensation factors are calculated based on both image data and respiratory waveform data, and these factors are continuously refined to optimize correction efficiency. The system uses the acquired image data itself to validate and adjust the motion compensation, ensuring high efficiency even when physiological motion is unstable.
Data Source
AI summary
The present disclosure provides a system and method for motion field generation and image correction. The method may include obtaining a plurality of first sets of magnetic resonance (MR) image data of an object generated based on a plurality of first sets of imaging sequences. The method may include obtaining a motion curve of the object. The method may include obtaining position emission tomography (PET) image data of the object generated in a scanning time period. The method may include generating one or more target motion fields corresponding to the scanning time period based on the plurality of first sets of MR image data and the motion curve. The method may include generating one or more corrected PET images by correcting, based on the one or more target motion fields, the PET image data.