Image Reconstruction Artifact Removal via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Reconstructed images in disease diagnosis often contain artifacts such as detector edge, projection object edge, and serrated artifacts, which can lead to misdiagnosis due to their presence in the images.
Innovation Solution
A method and system for image reconstruction that includes pre-processing operations like segmentation and geometrical error correction, followed by filtering and back projection to generate a reconstructed image, and an artifact removal module that specifically targets and removes detector edge, projection object edge, and serrated artifacts by determining their intersection points and dislocation information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional image reconstruction methods are used, then the reconstructed image can be generated quickly, but artifacts such as detector edge, projection object edge, and serrated artifacts appear in the image
Solution Approach 1:
The patent applies preliminary action by performing pre-processing operations on projection images before reconstruction. Specifically, it segments the projection image to identify regions containing artifacts, generates a negative film of the segmented image, and corrects geometrical errors in advance. This preliminary processing enables the reconstruction algorithm to operate more effectively on pre-cleaned data, reducing artifacts in the final reconstructed image while maintaining reasonable processing speed.
2Manufacturing precision
If artifact removal operations are performed on reconstructed images, then image quality improves, but processing time increases
Solution Approach 1:
The patent performs artifact removal operations in advance during the pre-processing stage rather than after reconstruction. By segmenting projection images to identify artifact-containing regions, generating negative films, and correcting geometrical errors before reconstruction, the system prevents artifacts from being fully formed in the reconstructed image. This approach reduces the need for time-consuming post-reconstruction artifact removal while maintaining image quality.
Solution Approach 2:
The patent extracts and isolates artifact-containing regions through image segmentation. By identifying and separating regions with detector edge artifacts, projection object edge artifacts, and serrated artifacts from the rest of the projection image, the system can apply targeted correction operations only to these problematic areas. This selective processing reduces overall computation time compared to processing the entire image.
3Manufacturing precision
If pre-processing operations including segmentation and geometrical error correction are applied, then artifact reduction is improved, but device complexity increases
Solution Approach 1:
The patent divides the projection image into multiple segments based on the presence and type of artifacts. By segmenting the image to identify regions containing detector edge artifacts, projection object edge artifacts, and serrated artifacts separately, the system can apply specific correction operations to each segment. This segmentation approach makes the complex artifact removal process more manageable and systematic.
Solution Approach 2:
The patent introduces a negative film as an intermediary element in the artifact removal process. The negative film is generated from the segmented projection image and serves as a mediator that facilitates geometrical error correction. This intermediary representation simplifies the correction process by providing a standardized format for applying geometrical transformations before final reconstruction.
Data Source
AI summary
A method and system for image reconstruction are provided. A projection image of a projection object may be obtained. A processed projection image may be generated based on the projection image through one or more pre-process operations. A reconstructed image including an artifact may be reconstructed based on the processed projection image. The artifact may be a detector edge artifact, a projection object edge artifact, and a serrated artifact. The detector edge artifact, the projection object edge artifact, and the serrated artifact may be removed from the reconstructed image.