On the efficiency of iterative ordered subset reconstruction algorithms for acceleration on GPUs

Abstract

Expectation Maximization (EM) and the Simultaneous Iterative Reconstruction Technique (SIRT) are two iterative computed tomography reconstruction algorithms often used when the data contain a high amount of statistical noise, have been acquired from a limited angular range, or have a limited number of views. A popular mechanism to increase the rate of convergence of these types of algorithms has been to perform the correctional updates within subsets of the projection data. This has given rise to the method of Ordered Subsets EM (OS-EM) and the Simultaneous Algebraic Reconstruction Technique (SART). Commodity graphics hardware (GPUs) has shown great promise to combat the high computational demands incurred by iterative reconstruction algorithms. However, we find that the special architecture and programming model of GPUs add extra constraints on the real-time performance of ordered subsets algorithms, counteracting the speedup benefits of smaller subsets observed on CPUs. This gives rise to new relationships governing the optimal number of subsets as well as relaxation factor settings for obtaining the smallest wall-clock time for reconstruction -a factor that is likely application-dependent. In this paper we study the generalization of SIRT into Ordered Subsets SIRT and show that this allows one to optimize the computational performance of GPU-accelerated iterative algebraic reconstruction methods.

Hybrid Imaging: Instrumentation and Data Processing

“…This is due to the numerous pre-processing and calibration steps in CT, which change the statistical properties of the measured data; the high spatial resolution in CT, which requires the application of edge-preserving regularization techniques; and the large amount of data that requires long reconstruction times. Nonetheless, the exponential growth of computer technology and the recent introduction of IR methods that can be implemented on a Graphic Processing Unit (GPU, [142]) render the use of iterative reconstruction in CT more and more a clinical commodity.…”

GPU-accelerated SART reconstruction using the CUDA programming environment

“…This trade-off between convergence and speedup is one of our main results. The convergence was recently examined by Xu et al 10 We compared reconstruction times on three different systems. First, an off-the-shelf PC equipped with an Intel Core2Duo processor running at 2 GHz, second, a workstation with two Intel Xeon QuadCore processors at 2.33 GHz and a NVIDIA QuadroFX 5600 with CUDA 1.1 and 2.0.…”

Evaluating popular non-linear image processing filters for their use in regularized iterative CT

“…Experiments have shown that the OS-SIRT ordered subsets scheme can increase the performance in both quality and speed [5]. We have used OS-SIRT here to demonstrate the improvements that can be obtained with nonlinear image processing filters forming the regularization mechanisms.…”