X-ray CT Image Reconstruction Z-Direction Filter Convolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional X-ray CT apparatuses face challenges in optimizing image reconstruction time and quality, particularly in controlling slice thickness during axial, cine, or helical scans, as existing methods either result in inconsistent image quality or prolonged reconstruction times.
Innovation Solution
The X-ray CT apparatus employs a combination of z-direction filter convolution in both projection data and image spaces, with adjustable coefficients based on slice sensitivity profiles and radiographic conditions, to optimize image reconstruction time and quality for various slice thicknesses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If z-direction filter convolution is performed in image space to control slice thickness, then image quality is improved, but image reconstruction time increases
Solution Approach 1:
The patent divides the filter convolution process into two segments: projection data space convolution (fast, controls slice thickness) and image space convolution (slower, refines image quality). By separating these functions, the system achieves both speed and quality without requiring the slower image space convolution alone, thus resolving the time-quality trade-off.
Solution Approach 2:
The projection data space z-direction filter convolution is performed as a preliminary action before image space convolution. This preliminary filtering controls the slice thickness and prepares the data for faster subsequent image space processing, reducing the overall reconstruction time while maintaining quality.
2Productivity
If slice thickness is increased to reduce reconstruction time, then processing speed improves, but image quality deteriorates
Solution Approach 1:
The patent dynamically adjusts the slice thickness control parameter (z-direction filter coefficients) based on the required image quality and reconstruction speed. The system can adaptively choose different convolution strategies depending on the clinical scenario, allowing optimization of both speed and quality rather than being fixed to a single approach.
3Loss of time
If conventional filter convolution methods are used, then processing time is reduced, but image quality becomes inconsistent
Solution Approach 1:
The patent incorporates feedback mechanisms where the image space convolution refines the results based on the projection data space convolution outcomes. This feedback loop ensures consistent image quality by adjusting and correcting artifacts introduced during the faster projection data processing, maintaining quality homogeneity across different slices and regions.
Data Source
AI summary
The present invention is intended to improve image quality ensured by a conventional (axial) scan, a cine scan, or a helical scan performed by an X-ray CT apparatus including a two-dimensional area X-ray detector that has a matrix structure. In helical scan image reconstruction based in three-dimensional image reconstruction performed in an X-ray CT apparatus including a two-dimensional area X-ray detector that is represented by a multi-array X-ray detector or a flat-panel X-ray detector and that has a matrix structure, an image expressing a slice thickness larger than the width of one detector array included in a multi-array X-ray detector is reconstructed according to either of a method of convoluting a z-direction filter to projection data items in the direction of detector arrays (z direction) and a method of convoluting a filer to a tomographic image space in the z direction. The two methods are optimized in terms of a calculation time and tomographic image quality. Consequently, a tomographic image can be quickly reconstructed with high quality.