基于GPU加速的锥束CT重建算法研究

锥束计算机断层扫描(Cone-Beam Computed Tomography,CBCT)具有采集速度快和空间分辨率高等特点,被生物医学等领域广泛关注。然而通过CPU串行处理CBCT重建中海量投影数据非常耗时,难以满足实时性的需求。GPU的发展为CBCT重建的并行加速提供了条件。根据三角函数周期性的特点对FDK算法进行了改进,并利用GPU实现了12幅投影数据同时并行计算。实验结果表明,相比于传统基于CPU的重建算法,基于GPU的CBCT重建算法在保证图像质量的前提下,将重建速度提高了超过310倍。

Research on Cone-Beam CT Reconstruction Algorithm Based on GPU Acceleration

With the characteristics of fast data acquisition and high spatial resolution, Cone-Beam Computed Tomography （CBCT） has been widely applied in biology and medical research. However, it is time-consuming to serially process the massive projection data in CBCT image reconstruction by CPU, which makes it difficult to satisfy the real-time requirement in industrial and clinical applications. The development of Graphics Processing Unit（GPU） facilitates parallel acceleration of CBCT reconstruction. This paper improves the reconstruction algorithm of FDK by greatly reducing the computation of trigonometric functions according to the characteristics of the periodicity of trigonometric functions, meanwhile, achieves the parallel calculation of 12 projections by GPU. The experimental results show that, compared with the traditional CPU-based FDK reconstruction algorithm, the proposed GPU-based CBCT reconstruction algorithm improves the reconstruction speed by more than 310 times while guaranteeing the quality of the reconstructed image.