MRI Acquisition Time Reduction via Atlas Regularization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional MRI systems incur significant acquisition and reconstruction time due to the need to create a regularization image from data acquired during the imaging process, which can be time-consuming and unnecessary for imaging healthy human brains that exhibit many similarities.
Innovation Solution
The system utilizes an atlas of stored MRI images to isolate and account for expected signal in under-sampled data sets, allowing for the removal of common features and reducing the data that needs to be acquired and reconstructed, thereby facilitating the use of pre-computed regularization images from a library of normal anatomy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a regularization image is created from data acquired during the imaging process, then the image quality is improved, but the acquisition time and reconstruction time increase
Solution Approach 1:
The system pre-acquires and stores regularization images from healthy subjects before the actual imaging procedure. These pre-acquired images serve as regularization references, eliminating the need to create regularization images during the scanning process, thereby reducing acquisition time while maintaining image quality
Solution Approach 2:
The system creates copies of regularization images from a library of pre-acquired healthy brain images. Instead of generating new regularization data during scanning, the system selects and uses appropriate copies from the library that match the current scan parameters, significantly reducing processing time
2Loss of information
If full data acquisition is performed, then the completeness of the data set is improved, but the acquisition time and data processing load increase
Solution Approach 1:
The system extracts only the essential information needed for reconstruction by using pre-acquired regularization images to represent common anatomical features. This allows the system to work with reduced data sets while maintaining completeness of critical information, thereby reducing acquisition time and processing load
Solution Approach 2:
The pre-acquired regularization images serve multiple functions: they provide anatomical reference, enable data reduction, facilitate image reconstruction, and allow for abnormality detection. This multi-functionality eliminates the need for separate processing steps, reducing overall acquisition and processing time
3Measurement precision
If conventional regularization processing is used, then the reconstruction accuracy is improved, but the reconstruction time and processing overhead increase
Solution Approach 1:
The system uses copied regularization images from a pre-computed library instead of generating new regularization data during reconstruction. This approach maintains reconstruction accuracy by using high-quality pre-processed references while significantly reducing the computational burden and reconstruction time
Solution Approach 2:
All regularization image processing, including normalization, registration, and quality assurance, is performed in advance during the pre-acquisition phase. This preliminary processing eliminates time-consuming operations from the reconstruction phase, improving productivity without sacrificing accuracy
Data Source
AI summary
Systems, methods, apparatus, and other embodiments associated with reducing imaging acquisition time are described. One example method includes accessing an under-sampled data set and a library of previously acquired data sets. The method includes producing an approximation of the under-sampled data set by transforming data stored in the library. The method includes producing a sparsified data set from the approximation and the under-sampled data set and then reconstructing the sparsified data set into a sparse image using a reconstruction technique configured to reconstruct sparse data. The method includes producing a fully-sampled approximation of the under-sampled data set and producing a final reconstructed image from the sparse image and the fully sampled approximation.