Motion Correction in Medical Imaging via Gaussian Weighted Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical imaging, particularly in breast MRI, patient movement between consecutive acquisitions introduces motion-related differences, making it difficult to track tissue characteristics like rapid contrast agent intake and washout, which are crucial for tumor detection, as prior art methods like optic-flow computation are inadequate for accurate motion correction.
Innovation Solution
A Gaussian weighted least mean square registration algorithm is used to derive a dense displacement field by processing feature maps from reduced-resolution images, warping the second image to correct for motion, with a multi-resolution strategy and iterative construction of displacement fields to ensure accurate alignment and robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If optic-flow computation is used for motion correction, then motion-related differences can be addressed, but measurement precision and reliability are insufficient for accurate tissue tracking
Solution Approach 1:
The patent changes the parameter space by computing optic flow not on original images but on Laplacian pyramid representations at multiple resolutions. This transformation of the image parameter space enables more reliable motion field estimation while maintaining precision through the multi-resolution approach and Gaussian weighting scheme
Solution Approach 2:
The patent segments the image processing into multiple resolution levels using a Laplacian pyramid. By dividing the problem into coarse-to-fine resolution stages, each level contributes to reliable motion estimation without sacrificing overall precision, as the fine levels refine the coarse-level motion fields
2Measurement precision
If multi-resolution strategy with iterative displacement field construction is used, then measurement precision and reliability improve, but device complexity and computational requirements increase
Solution Approach 1:
The algorithm segments the displacement field computation into iterative stages, where each iteration refines the field at a specific resolution level. This segmentation enables precise displacement estimation while managing complexity through systematic progression from coarse to fine details
Solution Approach 2:
The patent employs a dynamic, iterative construction of the displacement field rather than a static single-step computation. The displacement field evolves through multiple iterations and resolution levels, adapting progressively to achieve high precision while the iterative nature allows complexity to be managed in controlled stages
3Measurement precision
If dense displacement field is computed through Gaussian weighted least mean square optimization, then tissue characteristic tracking accuracy improves, but computational time and processing complexity increase
Solution Approach 1:
The computationally intensive dense displacement field computation is segmented across multiple resolution levels and iterative steps. By dividing the optimization problem into manageable stages using Laplacian pyramid representations, the patent achieves high tracking accuracy while reducing the computational burden of solving the full-resolution problem in a single step
Solution Approach 2:
The patent performs preliminary motion estimation at coarser resolution levels before computing the final dense displacement field at full resolution. This preliminary action at reduced complexity levels provides an initial solution that guides the subsequent fine-resolution optimization, achieving high accuracy with reduced overall processing time
Data Source
AI summary
A method for motion compensation between first and second images in a temporal sequence includes processing the first and second images in a reduction process for providing respective reduced resolution first and second images; deriving respective first and second feature maps from the respective reduced resolution first and second images, the feature maps including deriving the respective Laplacian of image data in the respective reduced resolution first and second images; deriving a displacement field by processing the first and second feature maps in accordance with a registration algorithm, the registration algorithm comprising solving, for each picture element or voxel, a local Gaussian weighted least mean square problem so as to derive respective vectors forming the displacement field; and warping the second image with the displacement field.