CT Truncation Artifact Correction via Iterative Boundary Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computed Tomography (CT) imaging systems face issues with truncation artifacts when patients extend beyond the scan field of view (SFOV) or are improperly aligned, leading to inaccurate image reconstruction due to unknown truncated data being set to zero.
Innovation Solution
An iterative method is employed to estimate and refine the patient boundary, classifying areas within and outside the SFOV as water and air, using forward projection and backprojection techniques to dilate or erode the boundary, and combining measured and estimated data to generate a corrected image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If truncated data is set to zero for image reconstruction, then the reconstruction process can proceed, but truncation artifacts appear as bright rings at the edge of the detector SFOV
Solution Approach 1:
The patent applies preliminary action by estimating the patient boundary and truncation region before performing the main image reconstruction. The system identifies the boundary between measured and truncated data regions in advance, then uses this information to guide the reconstruction process, thereby preventing artifacts rather than correcting them afterward.
Solution Approach 2:
The patent changes the parameter values in the truncated data region by estimating tissue types (water, bone, air) and assigning appropriate attenuation coefficients. Instead of using zero for truncated data, the system assigns physically meaningful values based on estimated tissue composition, which eliminates the bright ring artifacts while maintaining reconstruction accuracy.
2Manufacturing precision
If padding is used to set truncated data to non-zero values, then the brightness of the ring artifact is reduced, but the representation of truncated data outside the detector SFOV remains inaccurate
Solution Approach 1:
The patent introduces an intermediary estimation process that uses the measured data within the SFOV to infer properties of the truncated data outside the SFOV. By using the measured attenuation data and estimated patient boundary as intermediaries, the system can predict the characteristics of unmeasured regions without simply padding with arbitrary values.
Solution Approach 2:
The system employs feedback by iteratively refining the boundary estimate and tissue classification using the measured data. The estimated boundary and tissue types are used to generate expected projection data, which is compared with actual measured data, and the estimates are adjusted accordingly to improve accuracy.
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 effectively reduces truncation artifacts by improving the accuracy of image reconstruction, especially at the edges of the SFOV, providing a more accurate representation of the patient image.
Implementation Method 1
Computed Tomography (CT) imaging systems typically include an x-ray source and a detector. In operation, the x-rays are transmitted from the x-ray source, through a patient, and impinge upon the detector.
Data Source
AI summary
A method for performing truncation artifact correction includes acquiring a projection dataset of a patient, the projection dataset including measured data and truncated data, generating an initial estimate of a boundary between the measured data and the truncated data, using the measured data to revise the initial estimate of the boundary, estimating the truncated data using the revised estimate of the boundary, and using the measured data and the estimated truncated data to generate an image of the patient.