Bias Estimation for MRI Tissue Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical imaging techniques, particularly MR image analysis, face challenges in segmenting tissues due to large intensity bias and coil sensitivity-related spatial signal variations, leading to poor performance and tissue segmentation failures even with moderate shading.
Innovation Solution
A method and system that utilize synergies between anatomically matched PET and MR image data sets to estimate and correct spatial signal bias, employing a bias estimating and segmenting unit to generate baseline bias maps and body masks, enabling robust tissue segmentation resilient to image intensity artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If surface coil based image data acquisition is used to achieve higher encoding efficiency and high-resolution images, then productivity and image quality are improved, but large intensity bias and coil sensitivity-related spatial signal variations occur leading to poor segmentation performance
Solution Approach 1:
The patent introduces an intermediary bias estimation process that uses anatomically matched PET and MR image data sets to generate a bias map. This bias map serves as a mediator that corrects the intensity inhomogeneities in the MR images before segmentation is performed, thereby enabling accurate segmentation despite the use of surface coils with large intensity bias
Solution Approach 2:
The patent changes the parameter space by utilizing multi-contrast MR images (T1-weighted, T2-weighted, and proton density images) and combining them with PET data. By processing these multiple parameters together through the bias estimation algorithm, the system can separate true tissue signal variations from coil sensitivity-related intensity bias
2Reliability
If phase-field based tissue classification is used to achieve closed contour solution and noise resilience, then reliability is improved, but the method needs to be retuned to account for non-homogenous signal intensity distribution
Solution Approach 1:
The patent creates a universal bias estimation framework that works across different coil configurations and imaging protocols. The method uses anatomically matched PET and MR data together with a unified energy minimization approach that automatically adapts to different signal intensity distributions without requiring manual retuning of segmentation parameters
Solution Approach 2:
The patent performs preliminary bias estimation and correction before the actual segmentation process. By pre-computing the bias map from multi-contrast MR and PET images, the method eliminates the need for retuning phase-field parameters to account for intensity inhomogeneities, as the bias is already corrected in the pre-processed images
3Measurement precision
If body coil based image data acquisition is used to mitigate intensity inhomogeneity, then measurement precision is improved, but encoding efficiency decreases and high-resolution acquisition becomes less efficient
Solution Approach 1:
The patent merges the advantages of both body coil and surface coil approaches by combining multi-contrast MR imaging with PET imaging. The body coil provides uniform signal coverage while surface coils provide high spatial resolution and encoding efficiency. By fusing these complementary data sources and using the PET data to guide bias correction, the method achieves both uniformity and high resolution
Data Source
AI summary
A system and method for estimating image intensity bias and segmentation tissues is presented. The system and method includes obtaining a first image data set and at least a second image data set, wherein the first and second image data sets are representative of an anatomical region in a subject of interest. Furthermore, the system and method includes generating a baseline bias map by processing the first image data set. The system and method also includes determining a baseline body mask by processing the second image data set. In addition, the system and method includes estimating a bias map corresponding to a sub-region in the anatomical region based on the baseline body mask. Moreover, the system and method includes segmenting one or more tissues in the anatomical region based on the bias map.