CT Image Reconstruction Filtering for Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CT imaging systems face challenges in achieving high-quality image reconstruction due to limitations in memory and computational time, leading to artifacts such as jagged edges and undershoots, especially when using multi-material correction techniques with large reconstruction fields of view.
Innovation Solution
The method involves acquiring and reconstructing CT data, performing material characterization, forward projection, and error projection, with filtering applied at specific stages such as after material characterization or forward projection to reduce artifacts, using filters like low-pass filters to remove high-frequency information and edge sharpening techniques to improve image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the reconstruction size (pixel matrix) is increased to maintain high pixel resolution with a large DFOV, then image quality and material identification improve, but memory requirements and computational time increase significantly
Solution Approach 1:
The patent segments the reconstruction process into multiple passes: a first pass with a smaller pixel matrix (320×320) for speed, and a second pass with a larger pixel matrix (512×512) for final high-quality reconstruction. This segmentation allows the system to balance computational efficiency with image quality by performing material identification in the first pass and artifact correction in the second pass.
Solution Approach 2:
The patent performs preliminary material characterization and identification in the first reconstruction pass before the final image reconstruction. By identifying materials and their locations in advance using the smaller pixel matrix, the system can then apply targeted corrections in the second pass without needing to process the entire high-resolution reconstruction from scratch.
2Measurement precision
If the reconstruction size (pixel matrix) is increased to maintain high pixel resolution with a large DFOV, then image quality and material identification improve, but computational time increases significantly
Solution Approach 1:
The reconstruction process is divided into two sequential passes with different pixel matrix sizes. The first pass uses 320×320 for rapid material identification, and the second pass uses 512×512 for final high-quality reconstruction. This segmentation reduces total computational time compared to performing only high-resolution reconstruction.
Solution Approach 2:
Material characterization is performed in advance during the first reconstruction pass with the smaller pixel matrix. This preliminary action allows the second pass to focus computational resources on artifact correction and final reconstruction rather than repeating material identification, thereby reducing overall computational time.
3Productivity
If a small pixel matrix is used to accelerate reconstruction speed, then computational time decreases, but spatial resolution deteriorates causing artifacts in corrected images
Solution Approach 1:
The patent uses different pixel matrix sizes for different reconstruction passes: 320×320 for the first pass to achieve fast reconstruction speed, and 512×512 for the second pass to achieve high spatial resolution and eliminate artifacts. This segmentation allows each pass to be optimized for its specific purpose.
Solution Approach 2:
The first reconstruction image with the smaller pixel matrix serves as an intermediary product that enables rapid material identification. This intermediary reconstruction allows the system to then perform accurate artifact correction in the second pass without suffering from the resolution limitations of the smaller matrix throughout the entire process.
4Manufacturing precision
If filtering is applied to reduce artifacts, then image quality improves, but high-frequency information may be lost
Solution Approach 1:
Artifact correction is performed as a preliminary action in the second reconstruction pass after material identification is complete. By identifying materials and their locations in advance, the system can apply filtering selectively only in regions containing those materials, rather than applying global filtering that would remove high-frequency information from the entire image.
Solution Approach 2:
The patent applies filtering locally only in regions containing identified materials rather than globally across the entire image. This local quality approach allows artifact reduction in problem areas while preserving high-frequency information and sharp edges in regions without materials, thereby maintaining overall image quality without excessive information loss.
Data Source
AI summary
A method is provided including acquiring imaging data of an object to be imaged from a computed tomography (CT) detector. The method also includes reconstructing the acquired imaging data into an initial reconstruction image, and performing material characterization of an image volume of the initial reconstruction image to provide a re-mapped image volume. Further, the method includes performing forward projection on the re-mapped image volume to provide forward projection data, and providing an error projection based on the forward projection data. Also, the method includes filtering at least one of the initial reconstruction image, the re-mapped image volume, the forward projection data, or the error projection. The method also includes using the error projection to update the initial reconstruction image to provide an updated reconstruction image.