A Study of Iterative Reconstruction Algorithm Based on Penetrating Length Weighted for Computed Tomography

程明渊,贺奇才,胡琴明,陈弟虎
DOI: https://doi.org/10.3969/j.issn.1005-202x.2013.02.012
2013-01-01
Abstract:Objective: In traditional iterative reconstruction algorithm for computed tomography,to simplify the computation,the value of the matrix element of projection coefficient is determined by whether the X-ray path is through the pixel.If through the value is 1,else then the value is 0.Some rays that only go through the edge of the pixel are still assigned the value 1,which will enlarge the contribution of the ray projection.To reduce the image reconstruction error,and improve the quality of reconstruction image,a new iterative reconstruction algorithm based on penetrating-length weighted is introduced.By precise modeling it computes the value of the matrix element of the projection coefficient through the penetrating-length of the ray path within each pixel.Methods: Using the simulation tool of MATLAB7.0,scan the Shepp-Logan model and do the simulation of ML-EM iterative reconstruction,respectively based on the traditional projection coefficient and the penetrating-length weighted projection coefficient.Results: The simulation data shows that ML-EM iterative reconstruction based on penetrating-length weighted is better than ML-EM iterative reconstruction algorithm based on traditional projection coefficient,and it can improve the quality of reconstruction image.Conclusions: The ML-EM iterative reconstruction algorithm based on penetrating-length weighted can reduce image reconstruction errors.The precise modeling on the geometrical and physical factors for imaging caneffectively control the non-random portion to reduce the reconstruction error,and on one hand introduces a new method to compute the projection coefficient for iterative reconstruction algorithm,on the other hand,further improves the quality of CT reconstruction image in iterative algorithm.
What problem does this paper attempt to address?