MRI Image Reconstruction Using Time-Space Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current time-space parallel MRI techniques face challenges in accurately calculating the time-space correlation coefficient due to noise in data, leading to aliasing artifacts and reduced image accuracy when imaging moving objects.
Innovation Solution
The MRI apparatus and method involve a data processor that performs undersampling of MR signals from multiple coil channels to acquire undersampled K-t space data, and an image processor that calculates a time-space correlation coefficient based on noise information to restore noise-removed images, using a Kalman filter to correct line data and estimate unacquired data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If GRAPPA technique is used to process acquired MR signal, then image reconstruction is achieved, but aliasing artifacts and amplified noise occur when data is damaged or spatial interaction value changes due to noise
Solution Approach 1:
The patent introduces a time-space correlation coefficient as an intermediary parameter to model the relationship between adjacent K-space lines. This coefficient serves as a mediator that enables accurate signal estimation without directly relying on potentially damaged data, thereby reducing aliasing artifacts and amplified noise while maintaining image reconstruction accuracy.
Solution Approach 2:
The patent implements a feedback mechanism where the time-space correlation coefficient is calculated from actual measured data and used to estimate unmeasured signals. This feedback loop allows the system to continuously refine its signal estimation based on real measurements, improving reconstruction accuracy while suppressing artifacts and noise.
2Productivity
If time-space parallel MRI techniques are used to image moving objects, then time-serial imaging is achieved, but accuracy of time-space correlation coefficient is reduced due to noise in data
Solution Approach 1:
The patent changes the parameter used for correlation calculation from traditional spatial correlation to time-space correlation that incorporates temporal information. By adjusting the correlation model to account for time-serial imaging characteristics and using noise information from multiple coil channels, the system maintains high accuracy of the time-space correlation coefficient while preserving time-serial imaging capability.
Solution Approach 2:
The patent makes the correlation coefficient calculation universal by incorporating noise information from multiple coil channels into the correlation estimation process. This multi-functional approach allows the same correlation mechanism to work across different imaging scenarios and coil configurations, maintaining accuracy while enabling time-serial imaging.
3Speed
If undersampling is performed on MR signals from multiple coil channels, then acquisition speed is improved, but noise and aliasing artifacts increase in the restored image
Solution Approach 1:
The patent segments the K-space data into multiple coil channels and processes each channel separately through undersampling and correlation coefficient calculation. This segmentation allows independent optimization of each channel's contribution, enabling aggressive undersampling for speed improvement while controlling noise and artifacts through channel-specific correlation modeling.
Solution Approach 2:
The patent uses multiple coil channels as copies of the same signal with different noise characteristics. By leveraging the redundancy provided by multiple copies and calculating correlation coefficients based on noise information from all channels, the system achieves fast undersampled acquisition while suppressing noise and artifacts through the combined information from multiple copies.
Data Source
AI summary
The MRI apparatus includes a data processor, which time-serially performs undersampling on MR signals respectively received by coil channels included in a radio frequency (RF) multi-coil to acquire undersampled K-t space data, and an image processor that acquires a time-space correlation coefficient, based on noise information of the coil channels, and restores pieces of unacquired line data from the undersampled K-t space data by using the time-space correlation coefficient to acquire restored K-t space data, thereby increasing an accuracy of the time-space correlation coefficient to improve a quality of an image.