Dynamic Thread Configuration for CT Image Reconstruction Speed
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CT image reconstruction systems cannot achieve the fastest image reconstruction speed adaptively under different scanning and image-construction parameters due to fixed thread configurations, which limits hardware performance utilization.
Innovation Solution
A method and apparatus that dynamically generate thread configurations based on scanning data and image-construction parameters to optimize the number of threads in image reconstruction execution units, analyzing image reconstruction speeds to identify the optimal thread configuration for achieving the fastest speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the number of threads in execution units is increased to improve image reconstruction speed, then the image reconstruction speed increases, but the hardware resource consumption (CPU, memory, graphics card) increases and reaches a limitation
Solution Approach 1:
The patent implements dynamic thread configuration where the number of threads in execution units is adjusted based on actual image-construction parameters and scanning conditions. The system generates multiple thread configurations, evaluates their performance, and adaptively selects the optimal configuration, transforming the static thread setup into a dynamic, parameter-adaptive system that optimizes productivity without permanently increasing hardware resource consumption.
Solution Approach 2:
The system changes the parameter of thread configuration by generating multiple different thread configurations (varying the number of threads in different execution units) and selecting the optimal one based on evaluated image reconstruction speeds. This parameter optimization allows achieving faster reconstruction speeds by better utilizing existing hardware resources rather than simply increasing thread count indefinitely.
2Ease of operation
If a fixed number of threads is configured beforehand in execution units, then the system is simple to operate, but it cannot achieve the fastest image reconstruction speed under different scanning and image-construction parameters
Solution Approach 1:
The system performs preliminary evaluation of multiple thread configurations before actual image reconstruction. By pre-generating and evaluating different thread configurations to determine the optimal one for given image-construction parameters, the system prepares the best configuration in advance, ensuring fast reconstruction speed without requiring complex real-time adjustments during operation.
Solution Approach 2:
The system implements a feedback mechanism where image reconstruction speeds are evaluated for different thread configurations, and this evaluation feedback is used to select the optimal thread configuration. The system reads image reconstruction speeds, determines whether they meet preset conditions, and uses this feedback to identify and apply the optimal configuration, creating a closed-loop optimization process.
Data Source
AI summary
Disclosed are a method and apparatus for improving image reconstruction speed, to improve the image reconstruction speed by acquiring optimal thread configurations of execution units under different scanning conditions. The method includes: acquiring scanning data and extracting an image-construction parameter from the scanning data; generating thread configurations each comprising the numbers of threads in respective image reconstruction execution units, and obtaining image reconstruction speeds corresponding to the thread configurations by performing image reconstruction on the scanning data with the thread configurations; and obtaining image reconstruction speeds meeting a preset condition, grouping the thread configurations, which correspond to the image reconstruction speeds meeting the preset condition, into a candidate set, and extracting a thread configuration from the candidate set as an optimal thread configuration to be used in image reconstruction of the scanning data having the image-construction parameter.