Convolutional Spline Forward Back Projection for CT Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current iterative reconstruction algorithms for biomedical imaging, such as CT, PET, and SPECT, are hindered by high computational costs, leading to long reconstruction times and making them unsuitable for clinical use, despite their potential for low-dose imaging that reduces patient radiation exposure.
Innovation Solution
The implementation of a convolutional spline process for forward and back projection techniques, utilizing a box spline basis set for footprint and detector blur computation, which enables faster and more accurate image reconstruction from low-dose, few-view projection data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If iterative reconstruction algorithms are used to reduce patient radiation exposure, then image quality is improved, but reconstruction time increases significantly
Solution Approach 1:
The patent segments the projection data processing into discrete view subsets, where each view subset is processed independently through the iterative reconstruction algorithm. This segmentation allows parallel processing of multiple view subsets simultaneously, reducing overall reconstruction time while maintaining the low-dose imaging benefits of iterative reconstruction
Solution Approach 2:
The patent performs preliminary actions by pre-calculating system matrices and organizing projection data into structured view subsets before initiating the iterative reconstruction process. This preliminary organization enables more efficient computation during the actual reconstruction phase, balancing the trade-off between image quality and reconstruction time
2Object-affected harmful factors
If iterative reconstruction algorithms are implemented, then low-dose imaging capability is improved, but computational cost increases
Solution Approach 1:
By dividing the computational workload into separate view subset processing tasks, the patent enables parallel computation across multiple processors or GPU cores. This segmentation reduces the sequential computational burden while achieving the same low-dose imaging quality through iterative reconstruction
Solution Approach 2:
The patent applies partial action by performing iterative reconstruction on selected view subsets rather than processing all projection data through the full iterative algorithm. This selective approach reduces computational cost while still achieving sufficient image quality for low-dose imaging applications
3Productivity
If classical FBP algorithm is used for fast reconstruction, then reconstruction time is reduced, but radiation dose requirement increases
Solution Approach 1:
The patent combines the speed of FBP with iterative reconstruction by segmenting the data processing: FBP is applied to each view subset to generate initial images quickly, then iterative refinement is applied selectively. This hybrid approach maintains fast reconstruction speeds while enabling low-dose imaging capability
Solution Approach 2:
The patent uses FBP-reconstructed images as an intermediary step to provide initial estimates for the iterative reconstruction algorithm. This intermediary approach allows the system to leverage the speed of FBP while still achieving the low-dose imaging benefits through subsequent iterative refinement
Data Source
AI summary
An image is reconstructed an image from projection data by an image reconstruction computing entity. Projection data captured by an imaging device is received. An image is reconstructed and/or generated based at least in part on the projection data using a forward and back projection technique employing a convolutional spline process. The image is provided such that a user computing entity receives the image. The user computing entity is configured to display the image via a user interface thereof.