high performance computing on graphics processing units: hgpu.org

Medical Imaging is a technique aimed to develop tools to improve medical diagnosis and procedures. Medical imaging applications use different types of signals to create images of the inside of the human body. It is a very important technique that decreases the complication rate during procedures and provides with more information for more accurate diagnosis. One characteristic of image processing algorithms is that there are large data sets to be processed in order to provide visualization. In consequence, some diagnose imaging techniques take too long before their results are ready and the doctor can interpret them. A different approach can be used to decrease the time consumed by medical imaging algorithms. In this thesis, it has been explored the possibility of using Graphic Unit Processors (GPUs) for implementing real time medical imaging applications, more specifically ultrasound imaging for catheter ablation. The reader can find in the report the analysis of a set of medical imaging algorithms and its implementation on a GPU. There is a comparison with a CPU implementation and conclusions about the benefits and implications of developing a medical imaging application on top of a GPU.