Tomographic Reconstruction Using Hilbert Filtering and 2D Laplacian
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image reconstruction methods in 2D parallel-beam geometry, such as Filtered Backprojection (FBP), face limitations in achieving accurate and detailed 3D reconstructions, particularly in medical imaging applications like rotational angiography, where they struggle to provide clear low-contrast visualization of soft tissues.
Innovation Solution
The method involves applying Hilbert filtering to projection data, computing the antiderivative, backprojecting it into the image domain, and calculating the 2D Laplacian of the backprojection image, enhancing image reconstruction quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If Filtered Backprojection (FBP) is used for image reconstruction, then the reconstruction process is simple and fast, but the accuracy and clarity of low-contrast soft tissue visualization is insufficient
Solution Approach 1:
The patent changes the filtering parameter from conventional ramp filtering to Hilbert filtering, and introduces additional processing steps (computing antiderivative, applying 2D Laplacian) to modify the reconstruction formula. These parameter changes transform the FBP algorithm into a new method that achieves both speed and improved accuracy for low-contrast soft tissue visualization.
Solution Approach 2:
The patent replaces the conventional mechanical filtering operation with Hilbert filtering, which is a mathematical transformation that provides superior frequency response. This substitution enables better separation of signal and noise, improving the visualization of low-contrast structures while maintaining computational efficiency.
2Ease of operation
If conventional FBP algorithm is used, then the reconstruction process is straightforward, but the clarity of vascular anatomy and soft tissue visualization is poor
Solution Approach 1:
The patent segments the reconstruction process into distinct steps: Hilbert filtering of projection data, computation of antiderivative, backprojection, and application of 2D Laplacian. This segmentation allows each step to be optimized independently, maintaining operational simplicity while significantly improving soft tissue visualization clarity through the cumulative effect of these specialized operations.
Solution Approach 2:
The patent introduces an intermediary computation (antiderivative) between filtering and backprojection. This intermediary step acts as a mediator that transforms the filtered projection data into a form that, when backprojected and processed through the 2D Laplacian, yields clear soft tissue visualization. The intermediary computation enables the system to achieve high visualization clarity without complicating the overall operation.
3Volume of moving object
If 3D reconstructions are assembled from 2D slices, then volumetric imaging is achieved, but the detail and accuracy of the final 3D structure is limited
Solution Approach 1:
The patent applies the improved reconstruction method to each 2D slice independently, then assembles these enhanced 2D slices into a 3D volume. By operating in the 2D domain with the new Hilbert-filtering-based algorithm and then extending to 3D, the method achieves high accuracy in the third dimension while maintaining the computational advantages of 2D processing. The dimensional transition preserves and enhances the accuracy improvements from the 2D reconstruction technique.
Data Source
AI summary
An alternative analytical method for tomographic reconstruction in the 2D parallel-beam geometry is presented. This method may follow a filtering and backprojection scheme and may involve a global filtering in the projection domain and a local filtering in the image domain. For example, the method may include applying Hilbert filtering to the received projection data, computing an antiderivative of the filtered data, backprojecting the antiderivative into the image domain, and computing the 2D Laplacian of the backprojection image.