Acceleration Method of Back Projection for CT Image Reconstruction Based on Genetic Algorithm

Abstract

The invention discloses a CT image reconstruction back-projection acceleration method based on a genetic algorithm, which overcomes the problems of limited memory access bandwidth and low self-adaptive degree in the implementation mode of accelerating CT reconstruction back-projection by using CUDA in the prior art. Under the CUDA programming framework of this invention, the CUDA-based back-projection parallel computing strategy is as follows: (1) One back-projection kernel function operation processes multiple projections; (2) Reasonably design the thread grid (grid) and thread block (block) The dimensions and order of the global memory are guaranteed to be merged and accessed; (3) Use the constant memory to save the constant in the back projection; (4) Use the texture memory to realize bilinear interpolation and reduce the operation time of arithmetic instructions; (5) Reduce the back projection The same factor in the projection operation. The invention does not need to be identified manually, and the acceleration strategy has a high degree of self-adaptation.

Description

technical field [0001] The invention relates to a CT image reconstruction method, in particular to a genetic algorithm-based CT image reconstruction back-projection acceleration method. Background technique [0002] Computed Tomography (CT) has been widely used in the fields of industrial nondestructive testing and medical diagnosis due to its advantages of nondestructiveness, precision, and three-dimensional visualization. However, while CT equipment improves the imaging quality and reconstruction resolution, the corresponding computing resources and storage resources are very large. The reconstruction speed has become a bottleneck in the practical application of CT technology. How to accelerate the CT reconstruction process is currently the research field in this field. A difficult question. [0003] At present, the mainstream algorithms used in actual CT systems mainly focus on analytical reconstruction algorithms based on filtered back projection, such as FBP, FDK and o...

AI Technical Summary

Problems solved by technology

[0006] The present invention overcomes the problems of limited memory access bandwidth and low degree of self-adaptiveness in the realization method of accelerating CT reconstruction back-projection by using CUDA in the prior art, and provides a CT image reconstruction back-projection based on a genetic algorithm with a high degree of self-adaption Acceleration method

Images