CT Image Reconstruction Voxel Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Image reconstruction in CT technologies faces challenges with calculation speed and introduces artifacts, particularly due to the contribution of voxels depending on their location in the coordinate system of the detector, which affects image quality.
Innovation Solution
A method and system for image reconstruction that identifies and processes voxels differently based on geometric parameters, dividing them into sub-voxels and using iterative reconstruction processes, where voxels with specific footprint shadow ranges are calculated differently, and voxels representing air or couch areas are omitted or processed differently to enhance image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If iterative reconstruction process is used to improve image quality, then manufacturing precision is improved, but productivity deteriorates due to multiple iterations required
Solution Approach 1:
The image space is segmented into different voxel types (first kind and second kind) based on their geometric parameters and contribution factors. This segmentation allows different reconstruction strategies to be applied to different regions, enabling faster overall reconstruction while maintaining quality where needed.
Solution Approach 2:
Different reconstruction approaches are applied locally to different voxel types. High-contribution voxels (second kind) undergo full iterative reconstruction to ensure quality, while low-contribution voxels (first kind) use simplified methods. This local differentiation resolves the contradiction by applying computational effort only where it matters most for image quality.
2Manufacturing precision
If all voxels are processed uniformly with high calculation precision, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
Voxels are segmented into different categories based on their contribution factors and geometric parameters. This segmentation simplifies the overall calculation complexity by allowing different processing strategies for different voxel types, rather than applying a single complex algorithm to all voxels.
Solution Approach 2:
Full iterative reconstruction is applied only to voxels that significantly contribute to image quality (second kind voxels), while simplified methods are used for less important voxels. This partial application of complex processing reduces overall device complexity while maintaining reconstruction accuracy where it matters most.
3Productivity
If calculation time is reduced by skipping iterations, then productivity is improved, but manufacturing precision deteriorates due to reconstruction artifacts
Solution Approach 1:
The reconstruction process is segmented into different stages for different voxel types. Second kind voxels undergo multiple iterations to ensure accuracy, while first kind voxels use fewer iterations or simplified methods. This segmentation maintains image accuracy for critical regions while improving overall reconstruction speed.
Solution Approach 2:
Different iteration counts and reconstruction quality levels are applied locally to different voxel types based on their contribution factors. This ensures that regions with high impact on image quality receive充分的 iterative processing, while less critical regions use faster methods, resolving the speed-accuracy tradeoff.
Data Source
AI summary
The present disclosure relates to a system and method for generating an image. At least one processor, when executing instructions, may perform one or more of the following operations. When raw data relating to an object is retrieved, an image may be generated based thereon. A first voxel of the image is identified based a first geometric parameter relating to the first voxel; a second voxel of the image is identified based on a second geometric parameter relating to the second voxel; the image is reconstructed using an iterative reconstruction process, during which the calculation relating to the first voxel is based on the first number of sub-voxels, and the calculation relating to the second voxel is based on the second voxel.