X-ray Image Reconstruction Using Weighted Sinogram Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current X-ray CT imaging techniques face challenges in achieving optimal data interpolation for all pixel positions due to varying point response trajectories, leading to increased calculation loads and suboptimal image quality, especially when using methods like Fan-Para conversion.
Innovation Solution
The method involves generating representative sinograms through interpolation at multiple angles and performing weighted averaging based on the angle difference and trajectory slope for each pixel position during the image reconstruction process, reducing calculation loads while approximating ideal point response trajectory interpolation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If interpolation is performed along the point response trajectory for each pixel position to achieve optimal image quality, then image quality is improved, but calculation load becomes enormous and impractical
Solution Approach 1:
The patent segments the interpolation task by performing interpolation only at representative plural angle directions rather than for every possible pixel position. This creates a manageable subset of interpolation tasks that can be processed efficiently while still achieving good image quality through weighted averaging of the representative sinograms.
Solution Approach 2:
The patent applies different interpolation strategies to different regions by using weighted averaging where weights depend on the relationship between the representative angles and the actual trajectory slopes at each pixel position. This allows optimal interpolation characteristics to be applied locally where needed while avoiding unnecessary calculations elsewhere.
2Loss of information
If Fan-Para conversion is used to perform data interpolation, then data deficiency is addressed, but interpolation is only optimal for the central region of the DFOV and not for all pixel positions
Solution Approach 1:
The patent creates a universal interpolation solution that works for all pixel positions by generating representative sinograms at multiple angle directions and using weighted averaging. This single approach replaces the need for separate interpolation methods for different regions, providing consistent performance across the entire DFOV including the periphery where traditional Fan-Para conversion fails.
Data Source
AI summary
Provided is a technique in an X-ray imaging apparatus for implementing data interpolation approximating an ideal point response trajectory interpolation, along with reducing calculation load. A method is provided to perform interpolation at high speed, in a direction along the point response trajectory. Sinograms are interpolated in advance from measured data, as to representative angles (e.g., 0°, ±30°, ±60°, and 90°) only. When a pixel targeted for back projection is determined at the time of reconstruction, a slope of the point response trajectory is determined as to each view. According to the slope of the trajectory, each representative sinogram is added with weight, whereby interpolation data in association with any angle can be obtained.