Graphic Processing Unit (GPU) with Its Applications into the Hardware Acceleration of Three-Dimensional Medical Image Reconstruction

Recent growing development in Graphic Processing Units (GPUs) provides a new and low-cost solution to the digital signal processing, image processing or scientific problems that involve intensive and high-complexity computations. On the other hand, Computed Tomography (CT) has been a very important noninvasive tool in diagnostic medicine. In fact, the use of cone-beam projection based CT is growing in the clinical area due to its ability to provide three-dimensional (3D) information; also, the cone-beam CT can scan a wider area, in a shorter time, than the multi-slice CT. Rapid volumetric image reconstruction is of paramount importance to clinicians for prompt diagnosis and analysis of complex tissue alternations. However, due to a high demand on computation for either the analytical or iterative image reconstruction algorithms, it is almost impossible to meet the requirement of real-time reconstruction. This project aims to parallelize and implement a variety of 3D analytical and iterative image reconstruction algorithms on the GPU platform systems. Then, we will further evaluate and compare the execution performances obtained from the different parallelism strategies proposed in this project.

Project ID:PB10007-7277
External Project ID:NSC100-2221-E182-002