Fast GPU-Based CT Reconstruction using the Common Unified Device Architecture (CUDA)

Abstract

Abstract-The Common Unified Device Architecture (CUDA) is a fundamentally new programming approach making use of the unified shader design of the most current Graphics Processing Units (GPUs) from NVIDIA. The programming interface allows to implement an algorithm using standard C language and a few extensions without any knowledge about graphics programming using OpenGL, DirectX, and shading languages.We apply this revolutionary new technology to the FDK method, which solves the three-dimensional reconstruction task in cone-beam CT. The computational complexity of this algorithm prohibits its use for many medical applications without hardware acceleration. Today's GPUs with their high level of parallelism are cost-efficient processors for performing the FDK reconstruction according to medical requirements.In this paper, we present an innovative implementation of the most time-consuming parts of the FDK algorithm: filtering and back-projection. We also explain the required transformations to parallelize the algorithm for the CUDA architecture. Our implementation approach further allows to do an on-the-flyreconstruction, which means that the reconstruction is completed right after the end of data acquisition. This enables us to present the reconstructed volume to the physician in real-time, immediately after the last projection image has been acquired by the scanning device.Finally, we compare our results to our highly optimized FDK implementation on the Cell Broadband Engine Architecture (CBEA), both with respect to reconstruction speed and implementation effort.

Software architecture for multi-bed FDK-based reconstruction in X-ray CT scanners

“…A common strategy to speed up the back-projection step involves the use of graphics processing units (GPUs). In recent years several implementations have resulted in processing time reduction factors of up to 40 [9][10][11][12][13][14][15], most of them using the compute unified device architecture (CUDA). The modularity in our architecture facilitates the substitution of any module for a GPU kernel.…”

“…proposed by Feldkamp et al (FDK) [6] are still widely used for solving the 3D reconstruction task because of their straightforward implementation and computational efficiency [4]. Almost every aspect of the reconstruction process has been studied: there is literature on algorithm variations for different trajectories [7,8], optimizations using graphic processing units (GPUs) [9][10][11][12][13][14][15], strategies to reduce cone beam artifacts [16,17], study of consistency conditions [18], optimization of the back-projection step [19], etc. However, in a real practical system, the implementation of a reconstruction algorithm core such as FDK is just an initial step of the process, and there 1…”

Medical image processing on the GPU – Past, present and future

“…In 2007, the work by Xu and colleagues was further extended to real-time reconstruction (Xu and Mueller, 2007), such that interactive 3D image generation could be performed for flat-panel x-ray detectors and C-arm gantries. During the same year, Scherl et al (2007) were one of the first to use the CUDA programming language for FBP, while Yan et al (2008) still used OpenGL and Sharp et al (2007) used the Brook programming environment. Zhao et al (2009) focused on the problem of reconstruction of large volumes (e.g.…”

Cache-Aware GPU Optimization for Out-of-Core Cone Beam CT Reconstruction of High-Resolution Volumes

“…In particular, utilizing the compute unified device architecture (CUDA) compatible GPUs [12], to parallelize FDK computation has gained popularity due to its high-performance and low-cost implementation as compared to other devices [2], [13], [14]. In general, using implementations based on CUDA offloads the performance bottleneck of a CPU-based sequential code.…”