Virtual Non-Contrast X-Ray Image Registration for Motion Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
X-ray image registration is challenging due to patient movements and varying contrast agent concentrations, which cause registration problems in imaging techniques like CT systems, especially during perfusion and multiphase acquisitions, as different contrast agents lead to different CT values and dynamic changes over time.
Innovation Solution
A motion correction method that generates virtual non-contrast X-ray images through material decomposition of spectral raw X-ray data from different contrast distributions, allowing for registration of these images to determine a transformation field and correct for patient movement, using a decomposition unit and registration unit within an X-ray imaging system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If different contrast agents or different contrast-agent concentrations are used to visualize different types of tissue, then the diagnostic information and tissue differentiation are improved, but registration accuracy deteriorates due to patient movements and varying CT values
Solution Approach 1:
The patent introduces virtual non-contrast images as an intermediary representation that mediates between multiple contrast-enhanced images. These virtual non-contrast images serve as a common reference framework that is insensitive to contrast agent variations, enabling accurate registration while preserving the diagnostic information from different contrast phases. The virtual non-contrast images act as a mediator that translates between different contrast representations.
Solution Approach 2:
The patent transforms the image data by changing the contrast parameter through material decomposition. By decomposing the X-ray attenuation into material-specific components (bone, soft tissue, contrast agent), the system can reconstruct images with standardized contrast behavior, effectively changing the contrast parameter from variable (different contrast agents/concentrations) to controlled (virtual non-contrast representation).
2Ease of manufacture
If conventional motion correction algorithms using normalized mutual information are used, then the method is simple and computationally efficient, but registration accuracy deteriorates when contrast agent concentrations vary
Solution Approach 1:
The patent performs preliminary transformation of the contrast-enhanced images into virtual non-contrast images before applying registration. This preliminary action removes the problematic variable (contrast agent concentration) from the images, creating a standardized representation that enables accurate registration. By preprocessing the images to eliminate contrast variations, the subsequent registration can use simple algorithms effectively.
Solution Approach 2:
The patent creates virtual copies of the original images in a standardized representation (virtual non-contrast images). These copied representations maintain the anatomical structure and spatial information while removing the variable contrast characteristics, allowing registration to be performed on the copies rather than the original variable contrast images.
3Loss of time
If landmark detection is used to initialize motion field, then the registration process is accelerated, but robustness deteriorates when contrast agent concentrations change
Solution Approach 1:
The patent changes the contrast parameter of the images to virtual non-contrast representation, which provides consistent anatomical features regardless of contrast agent concentration. This parameter change ensures that landmarks detected in virtual non-contrast images remain robust and reliable across different contrast phases, as the anatomical structures are depicted with standardized contrast behavior.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate registration of X-ray images with different contrast agent concentrations, improving the alignment of image data and correcting for patient movement, thereby enhancing the reliability of image registration and maintaining consistent contrast behavior across images.
Implementation Method 1
a first virtual non-contrast X-ray image is generated. The first virtual non-contrast X-ray image is determined on the basis of first spectral raw X-ray data by way of material decomposition
Data Source
AI summary
In an embodiment of a motion correction method, a first virtual non-contrast X-ray image of a region under examination is determined based upon first spectral raw X-ray data associated with a first contrast distribution, via material decomposition. In addition, a second virtual non-contrast X-ray image of the region under examination is determined based upon second spectral raw X-ray data associated with a second contrast distribution, differing from the first contrast distribution, via material decomposition. Then the first virtual non-contrast X-ray image is registered with the second virtual non-contrast X-ray image to determine a transformation field between the two virtual non-contrast X-ray images. Finally, based upon the determined transformation field, first X-ray image data based on the first raw X-ray data is registered with second X-ray image data based on the second raw X-ray data. An X-ray imaging method, a motion correction device and an X-ray imaging system are also discussed.