RF Magnetic Field Mapping via Regression Analysis in MRI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical imaging techniques, such as MRI, face challenges in accurately and efficiently determining physical quantities like radiofrequency magnetic fields on biological tissues, particularly in ultrahigh field MRI where inhomogeneities lead to spatially dependent inaccuracies, requiring longer acquisition times or compromising on accuracy.
Innovation Solution
A computing system that digitizes biological tissues into voxels or pixels, spatially correlates them, and uses regression analysis to generate accurate mappings of physical quantities, leveraging machine learning algorithms and cloud computing for efficient data processing and calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If conventional FLASH imaging methods are used to reduce acquisition time, then acquisition time is reduced to seconds, but measurement precision deteriorates due to spatially dependent inaccuracies in ultrahigh field MRI
Solution Approach 1:
The patent introduces an intermediary computational processing step between the fast FLASH imaging acquisition and the final mapping result. A regression model is trained using ground truth data from accurate but slow AFI measurements, then applied to correct and enhance the accuracy of fast FLASH measurements. This intermediary model acts as a bridge, transferring knowledge from slow accurate measurements to fast inaccurate measurements, thereby improving precision without sacrificing speed.
Solution Approach 2:
The patent transforms the problem by changing the parameter representation of the RF magnetic field mapping. Instead of directly measuring the field, the system uses a regression model that learns the relationship between FLASH-derived parameters and actual B1+ field values. The model processes the fast-acquired data through a trained computational framework, transforming inaccurate measurements into accurate mappings by leveraging patterns learned from ground truth data.
2Measurement precision
If AFI method is used to improve measurement precision of RF magnetic field, then accuracy is improved, but loss of time increases to minutes due to lengthy acquisition time
Solution Approach 1:
The patent performs preliminary action by pre-training a regression model using accurate AFI measurements as ground truth data. This training phase captures the relationship between FLASH and AFI measurements. Once trained, the model can be applied to new FLASH data without requiring time-consuming AFI acquisitions, thus achieving accurate mappings through fast preprocessing and model application rather than slow direct measurement.
Solution Approach 2:
The patent creates a computational copy of the accurate AFI measurement process through the regression model. Instead of repeatedly performing slow AFI measurements, the system learns from a set of AFI ground truth data and creates a model that replicates the accurate measurement process. This computational copy can then be applied instantaneously to FLASH data, achieving the accuracy of AFI without its time cost.
3Measurement precision
If linear mapping is used to correct FLASH measurements, then some improvement in accuracy is achieved, but manufacturing precision remains insufficient because spatially dependent inaccuracies are not captured
Solution Approach 1:
The patent moves from simple linear mapping (one-dimensional correction) to a multi-dimensional regression approach. The regression model considers multiple input features and spatial relationships, transforming the correction process into a higher-dimensional computational space. This allows the model to capture complex spatially dependent patterns that linear methods cannot, thereby improving both measurement precision and spatial accuracy of the mapping.
Data Source
AI summary
A computing system for providing a mapping of a physical quantity on a biological tissue includes a data interface configured to obtain data from the physical quantity on different spatial points of the biological tissue. The computing system includes a computation module having: a digitization unit configured to produce a digitized representation of the biological tissue in voxels and/or pixels; a concatenation unit configured to spatially correlate the voxels and/or pixels of the produced digitized representation of the biological tissue; and a regression unit configured to process the information of the spatially correlated voxels and/or pixels and generate a regression analysis of the physical quantity on the biological tissue. The computing system includes an output data interface configured to, based on the generated regression analysis, provide a mapping of the physical quantity on the biological tissue. A value of the physical quantity is assigned to each voxel and/or pixel.