Tomography Reconstruction Filter for Noise and Edge Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current filtered backprojection methods in tomography fail to adequately account for noise and edge blurring in reconstructed images, leading to suboptimal image clarity and noise propagation.
Innovation Solution
A system and method that apply a filter to intermediate images based on a noise model and an image model, using weights generated from the noise model to preserve edges and reduce noise propagation, while merging angular information with image spatial features in an intermediate space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional reconstruction filters (derivative or ramp filters) are applied in filtered backprojection, then the reconstruction process can be completed, but noise and edge blurring are not adequately accounted for, leading to poor image quality
Solution Approach 1:
The patent changes the parameters of the filtering process by applying different filters to different components of the backprojected image. Specifically, a first filter is applied to edge-aligned components while a second filter is applied to non-edge components, allowing optimization of both edge preservation and noise reduction simultaneously
Solution Approach 2:
The patent applies local quality by treating different regions of the image differently based on their characteristics. Edge regions receive filtering optimized for preserving sharp transitions, while non-edge regions receive filtering optimized for noise reduction, thereby improving overall image quality without compromising critical features
2Manufacturing precision
If post-filters are applied to reconstructed images to improve edge clarity, then edge clarity may be improved, but noise within the projections is not adequately accounted for
Solution Approach 1:
The patent performs preliminary action by separating the image into edge-aligned and non-edge components before applying filters. This allows the filtering process to address both noise and edge clarity simultaneously from the beginning, rather than attempting to correct edge clarity after noise has already degraded the image
3Object-affected harmful factors
If pre-filters are applied to projections prior to reconstruction to account for noise, then noise may be reduced, but edge clarity in the reconstructed image is not adequately improved
Solution Approach 1:
The patent applies segmentation by dividing the backprojected image into distinct components: edge-aligned components and non-edge components. This segmentation allows each component to be filtered independently with appropriate filters, thereby simultaneously achieving noise reduction and edge clarity improvement that cannot be achieved with a single uniform filter
Data Source
AI summary
A method for reconstructing an object via tomography is provided. The method includes: acquiring a plurality of projections of the object at different angles via an imaging system; backprojecting each projection of the plurality to generate an intermediate image of each projection via at least one processor of the imaging system; and applying a filter to each intermediate image via the at least one processor. The filter is based at least in part on a noise model and an image model.