CT Surface Mesh Optimization via Projection Density Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating surface data from computer tomographic measurements often result in stair artifacts, which distort flat material boundaries into staircase structures due to the voxel-based calculation model.
Innovation Solution
The method involves creating a volume model from projection images, generating an initial surface mesh, and optimizing surface nodes using auxiliary points defined from the projection images, allowing direct reconstruction of density values at these points to refine the surface mesh positions, thereby eliminating interpolation assumptions and improving precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional voxel-based methods are used to generate surface data from CT volume data, then the calculation process is simple and fast, but stair artifacts occur that distort flat material boundaries into staircase structures
Solution Approach 1:
The patent introduces auxiliary points positioned outside the surface mesh in three-dimensional space, adding spatial dimensions to the evaluation process. By evaluating density values at these external auxiliary points rather than only at mesh nodes, the method transcends the conventional voxel grid limitations and achieves sub-voxel surface precision without requiring complex iterative optimization of the mesh structure itself.
Solution Approach 2:
The patent uses auxiliary points as intermediary evaluation locations between the volume data and the surface mesh. These auxiliary points serve as mediators that capture density information from the projection data, allowing the surface nodes to be positioned with precision finer than the voxel grid without directly manipulating the voxel structure. This intermediary approach resolves the contradiction by decoupling surface precision from voxel resolution.
2Measurement precision
If the Marching Cube algorithm or similar voxel-based methods are used, then computational effort is low and processing is fast, but stair artifacts distort flat material boundaries
Solution Approach 1:
The patent performs preliminary evaluation of density values at auxiliary points before finalizing surface mesh node positions. By pre-calculating density values at strategically positioned auxiliary points outside the surface, the method prepares accurate reference data that guides node positioning, achieving high boundary detection accuracy without requiring time-consuming iterative optimization during the main processing phase.
Solution Approach 2:
The patent replaces the mechanical voxel-based marching cube algorithm with a projection-based density evaluation system. Instead of mechanically traversing voxel grids and applying thresholding rules that inherently produce stair artifacts, the method substitutes a continuous density evaluation approach using auxiliary points and projection data, achieving smoother and more accurate boundary representation with comparable computational efficiency.
3Manufacturing precision
If auxiliary points outside the surface mesh are used for evaluation, then sub-voxel precision is achieved and stair artifacts are reduced, but computing effort increases
Solution Approach 1:
The patent applies local quality by positioning auxiliary points specifically at locations outside the surface mesh where they can most effectively evaluate boundary accuracy. Rather than uniformly increasing computation throughout the entire volume, the method concentrates computational effort at strategic auxiliary point locations that provide maximum information for sub-voxel surface positioning, achieving high precision with minimal additional computational energy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively reduces stair artifacts by directly reconstructing density values from measurement data, providing more accurate surface representations with increased computing effort but improved precision.
Implementation Method 1
at least one radiation source (source of invasive radiation, for example an X-ray source)
Implementation Method 2
the invasive radiation from the radiation source is attenuated to varying degrees as it passes through the object
Implementation Method 3
The detector unit records data (projection images) that represent a two-dimensional image of the object
Implementation Method 4
Using a reconstruction algorithm, a volume data set can be calculated from the projection images
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for generating surface data of an object (2) using projection images, which have been recorded via computed tomography measuring of the object (2) positioned in the beam path (6) between a radiation source (3) and a detector unit (5), comprising the steps of I) creating a first volume model with the projection images; II) generating a first surface network (N) with node points based on the volume model; characterised by III) optimising the surfaces of the surface network (N) with the aid of projection images recorded by the detector unit (5).