CT Image Reconstruction Arithmetic Device Processing Time Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing successive approximation method for image reconstruction in CT examinations incurs a high calculation load due to matrix calculation processes like back projection and forward projection, leading to increased processing time.
Innovation Solution
An arithmetic device and method that creates object projection data, a reference image, and a standardized image with defined noise reduction coefficients, using repetitive processing to update images by calculating differences between reference and updated images, thereby omitting the high-load matrix calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the successive approximation method is used for image reconstruction, then noise reduction effect is improved, but processing time increases due to matrix calculation processes
Solution Approach 1:
The patent extracts and removes the computationally intensive matrix calculation processes (forward projection and back projection) from the successive approximation method, replacing them with simpler image processing operations. This extraction of the harmful computational burden maintains the noise reduction capability while significantly reducing processing time.
Solution Approach 2:
The patent substitutes the mechanical matrix calculation system with an image processing-based system. Instead of performing repeated forward and back projection matrix operations, the invention uses image space operations that achieve the same noise reduction effect with much lower computational complexity, thereby reducing processing time while maintaining the noise reduction benefit.
2Measurement precision
If matrix calculation processes are performed for image update, then noise reduction is achieved, but calculation load increases
Solution Approach 1:
The patent extracts the matrix calculation processes from the image update operation and replaces them with image processing operations. This removes the computational burden of matrix calculations while preserving the essential noise reduction function through alternative image space operations.
Solution Approach 2:
The patent changes the operational parameters from matrix space calculations to image space operations. By transforming the problem domain from projection data space to image space, the calculation load is significantly reduced while maintaining the noise reduction effect through different computational approaches.
3Measurement precision
If forward projection and back projection processes are performed, then image reconstruction accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts and eliminates the forward projection and back projection processes from the reconstruction algorithm. These computationally intensive processes are replaced with direct image processing operations that achieve similar reconstruction accuracy without the time penalty of repeated matrix calculations.
Solution Approach 2:
The patent substitutes the projection-based mechanical calculation system with an image processing-based system. The forward and back projection operations are replaced with image space operations that maintain reconstruction accuracy while dramatically reducing processing time through simpler computational mechanisms.
Data Source
AI summary
In order to provide an arithmetic device, an X-ray CT apparatus, and an image reconstruction method, capable of reducing processing time while maintaining a noise reduction effect, in a successive approximation image reconstruction method (separable paraboloidal surrogate (SPS) method) of the related art, updated images are forward-projected, whenever images are repeatedly updated, a difference between forward projection data and original object projection data is back-projected so that a difference image is obtained, and a forward projection process and a back projection process are repeatedly performed, but, in the present invention, a forward projection process and a back projection process requiring calculation time are replaced with a process requiring a relatively small calculation amount, such as a difference between an updated image and a reference image, and, as a result, it is possible to considerably reduce a calculation amount in a successive approximation image reconstruction process and to reduce processing time.