Deformable Registration for 4D Cone Beam CT Image Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 4D cone beam CT systems face challenges in accurately reconstructing volumetric images of moving targets, such as those affected by breathing motion, as they require precise synchronization and correction of projection images across different phases of a physiological cycle, which can result in incomplete or noisy data due to limited radiation doses and motion artifacts.
Innovation Solution
The method involves sorting projection images based on breathing phases and using image registration techniques to modify and combine projection images from different phase bins, allowing for the construction of improved volumetric images that reduce noise and enhance image quality without increasing radiation dose, by 'borrowing' data from other phases and applying deformation registrations to create a composite image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If projection images are sorted into different phase bins and reconstructed separately, then volumetric images at different breathing states can be obtained, but the images contain noise and motion artifacts due to limited radiation doses per phase bin
Solution Approach 1:
The patent combines projection images from multiple phase bins to reconstruct a single volumetric image. By merging data from different phases, the system accumulates more projection images per reconstruction, improving signal-to-noise ratio and reducing artifacts while maintaining motion compensation through registration techniques.
Solution Approach 2:
The patent performs preliminary motion compensation and registration of projection images before reconstruction. By pre-aligning images from different phases using deformation registration, the system reduces motion artifacts in the final reconstruction without requiring higher radiation doses.
2Measurement precision
If deformation registration is applied to combine projection images from different phase bins, then noise is reduced and image quality is enhanced, but the computational complexity and processing time increase
Solution Approach 1:
The patent divides the deformation registration process into manageable segments: first sorting projection images into phase bins, then selectively combining bins, and finally applying registration only to the combined set. This segmentation reduces overall computational complexity compared to processing all images simultaneously.
Solution Approach 2:
The patent applies deformation registration selectively rather than to all projection images. By choosing to combine only certain phase bins and applying registration only to those combinations, the system achieves noise reduction while limiting computational overhead to necessary operations only.
3Object-affected harmful factors
If multiple projection images from different phase bins are combined, then radiation dose can be reduced while maintaining image quality, but the synchronization and sorting accuracy requirements increase
Solution Approach 1:
The patent uses breathing monitoring feedback to accurately sort projection images into phase bins. The system continuously monitors respiratory motion and uses this feedback to correctly assign each projection image to its corresponding phase bin, ensuring accurate synchronization even when combining images from multiple bins to reduce radiation dose.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A method of obtaining a volumetric image includes obtaining a plurality of volumetric images, the volumetric images generated using respective sets of projection images, wherein the volumetric images and the respective sets of projection images correspond with different respective bins for a physiological cycle, and determining an additional volumetric image using one or more of the projection images from two or more sets that correspond with different respective bins for the physiological cycle, wherein the act of determining the additional volumetric image is performed using a processor.