Static CT Sub-Image Reconstruction for Distributed Ray Sources
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The traditional image reconstruction methods used in slip ring CT apparatuses are not applicable to static CT apparatuses with distributed ray sources due to differences in arrangement and beam output modes, leading to challenges in obtaining high-quality CT images.
Innovation Solution
A method for static CT apparatuses involving a distributed ray source with multiple ray source points, using image segmentation, optimization with neural networks, and merging sub-images to enhance image quality, including pre-training neural networks to minimize projection data differences and iteratively refining the image reconstruction process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional image reconstruction methods are used for static CT apparatus, then the apparatus structure is simple, but the image quality is insufficient
Solution Approach 1:
The patent divides the CT image into multiple sub-images corresponding to different ray source points, processes each sub-image separately through neural networks, and then synthesizes them into a final high-quality image. This segmentation approach enables targeted optimization of each sub-image while maintaining overall image quality improvement.
Solution Approach 2:
The patent introduces neural networks as intermediary components between the raw projection data and the final CT image. These neural networks serve as mediators that learn the mapping relationship from projection data to image data, enabling high-quality image reconstruction without requiring complex traditional algorithms.
2Adaptability or versatility
If distributed ray source with multiple points is used, then scanning flexibility is improved, but image reconstruction difficulty increases
Solution Approach 1:
The patent segments the reconstruction task by dividing the CT image into multiple sub-images, each corresponding to a specific ray source point. This segmentation transforms the complex multi-source reconstruction problem into multiple simpler single-source reconstruction problems that can be processed independently and then combined.
Solution Approach 2:
The patent uses neural networks to learn and copy the mapping relationship from projection data to image data. By training the neural network on training datasets, the system copies the reconstruction knowledge from training examples to new input data, simplifying the reconstruction process for distributed ray sources.
3Manufacturing precision
If neural network optimization is applied to sub-images, then image resolution is improved, but computational load increases
Solution Approach 1:
The patent divides the image into multiple sub-images and processes each sub-image separately through neural networks. This segmentation reduces the computational complexity of processing the entire image at once, as each neural network only needs to handle a portion of the total image data.
Solution Approach 2:
The patent applies neural network optimization selectively to sub-images rather than processing the entire image uniformly. This partial action approach focuses computational resources on specific regions that benefit most from neural network optimization, reducing overall computational load while maintaining image quality.
Data Source
Figure 1~2A
Figure 2B~3A
Figure 3B
AI summary
A static CT apparatus and an imaging method for the same are provided. The imaging method includes: acquiring initial projection data of an inspected object at different angles by using a distributed ray source and a detector, where the initial projection data includes projection data that is directly obtained by the detector based on the rays emitted from a plurality of ray source points; obtaining a first CT image using a reconstruction algorithm according to the acquired initial projection data; dividing the first CT image into N first sub-images, where N is a positive integer greater than or equal to 1, and a union of the N first sub-images covers the entire first CT image; optimizing the N first sub-images to obtain N second sub-images; and merging the N second sub-images to obtain a second CT image.