Iterative CT Reconstruction Across Raw, Projection, and Image Domains
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current X-ray computed tomography (CT) systems rely on oversimplified mathematical and physical models for image reconstruction, which do not fully utilize the information obtained, leading to suboptimal CT image quality.
Innovation Solution
A system for iteratively reconstructing CT images through three domains: raw data, projection data, and image data, using a raw domain processor to generate serial projection data, a projection domain processor to generate serial image data, and an image domain processor to minimize a cost function, with iterative loops and polynomial functions to refine coefficients and image data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If oversimplified mathematical and physical models are used for image reconstruction, then the reconstruction process is computationally simple and fast, but the CT image quality and accuracy are suboptimal
Solution Approach 1:
The patent implements iterative feedback loops where the reconstruction process continuously refines the image by comparing the reconstructed projection data with the actual measured data. The image domain processor generates updated images, which are then back-projected to the projection domain, and discrepancies are used to update the raw data domain model parameters. This feedback mechanism allows the system to progressively improve image quality while converging to an optimal solution.
Solution Approach 2:
The patent introduces a multi-domain approach that operates across three distinct domains: raw data domain, projection data domain, and image data domain. By transforming the reconstruction problem across these different dimensional representations and utilizing polynomial functions to model the X-ray spectrum in the raw data domain, the system captures complex physical effects that single-domain methods miss, thereby improving image quality without requiring overly complex models in any single domain.
2Loss of information
If detailed and accurate physical models are used for image reconstruction, then the CT image quality and information utilization are improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent divides the reconstruction process into three distinct domain processors: raw domain processor, projection domain processor, and image domain processor. Each processor handles specific aspects of the reconstruction task and operates on data in its appropriate domain. This segmentation allows the system to process different types of information in parallel and reduces the computational burden on any single processor, thereby reducing overall processing time while maintaining comprehensive information utilization.
Solution Approach 2:
The patent employs polynomial functions to model the X-ray spectrum and uses iterative updates to refine model parameters across domains. By representing the complex physical model with polynomial parameters that can be efficiently updated through iterative optimization, the system captures detailed physical effects without requiring computationally expensive simulations at each iteration, thus balancing accuracy with processing speed.
3Measurement precision
If iterative reconstruction methods are used to fully utilize X-ray signal information, then the accuracy of tissue composition and density identification is improved, but the computational resources and processing complexity increase
Solution Approach 1:
The patent leverages the polynomial representation of the X-ray spectrum in the raw data domain as an additional dimensional approach. By modeling the energy-dependent attenuation characteristics using polynomial functions, the system extracts more information from the X-ray signals without requiring complex spectral measurements. This polynomial domain processing, combined with iterative refinement across three domains, enhances tissue identification accuracy while keeping the processing architecture manageable through mathematical transformations rather than hardware complexity.
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
This approach enhances the accuracy and quality of CT image reconstruction by fully utilizing the information from X-ray signals, improving the identification of tissue compositions and densities.
Implementation Method 1
X-ray CT systems, X-rays are used to image internal structures and features of a region of a subject or an object
Data Source
AI summary
A system for iteratively reconstructing computed tomography images through three domains is disclosed. The system comprises a raw domain processor, a projection domain processor, and an image domain processor. The system also comprises two iterative loops: one is through a raw synthesizer connecting the raw domain processor and the projection domain processor, and the other is through a projection synthesizer connecting the projection domain processor and the image domain processor respectively.