AN ACCELERATED ORDERED SUBSETS RECONSTRUCTION ALGORITHM FOR X-RAY CONE-BEAM CT
Iterative image reconstruction algorithms have many advantages over analytical image reconstruction algorithms.The OSEM (ordered subsets EM) iterative algorithm has enjoyed considerable interest for computed tomography due to its acceleration of the ML-EM algorithm.Compared with the conventional OSEM algorithm,the OS method RAMLA (row action ML algorithm) can not only bring about significant acceleration in the iterative reconstruction,but also outperform OSEM in convergence rate.In this paper,we apply 3D RAMLA algorithm to X-ray cone-beam CT image reconstruction.In order to further speed up the 3D RAMLA algorithm while still retaining its convergence properties,an improved RAMLA algorithm is proposed.By increasing the step size of the correction factor,the algorithm achieves a great deal of acceleration in convergence speed.The GPU-based implementation of this algorithm is also presented in this paper.The advantages of the method are verified by the experiment of the 3D image iterative reconstruction of computer-simulation data.
computed tomography cone-beam ordered subsets iterative algorithm
Xing Zhao Jing-jing Hu
School of Mathematical Sciences,Capital Normal University,Beijing 100048,China Department of Computer Science,Beijing Institute of Technology,Beijing 100081,China
国际会议
北京
英文
474-477
2011-10-19（万方平台首次上网日期，不代表论文的发表时间）