PET reconstruction on the GPU

The goal of this project is to develop a high performance evaluation and processing software library, which - due to the heavy parallelization - can greatly speed up 3D reconstruction methods, thus, they can be used in clinical routine works as well. By increasing the image quality using only software methods, our system can be used both in the newest imaging devices to exploit their capabilities and in previously installed devices to increase their efficiency. Since it is indispensable to validate our system before clinical use, our system goes through an extensive verification process. The benefits of increasing spatial resolution, contrast and signal/noise ratio come in multiple levels:

  • more accurate diagnosis, earlier recognition of illnesses
  • reduced radiation level due to the better signal/noise ratio
  • decreased examination time, increased patient throughput

Participants (from the department)

  • Staff: Benyó Balázs, Benyó Zoltán, Szlávecz Ákos, Szécsi László, Szirmay-Kalos László, Umenhoffer Tamás, Tóth Balázs
  • PhD students: Hesz Gábor, Jakab Gábor, Magdics Milán, Puskás Zoltán
  • MSc students: Barta Pál, Csendesi Ádám, Kovács Balázs, Márta Zsolt, Molnár Milán,

Partners in consortium

  • MEDISO Medical Imaging Systems
  • Budapest University of Technology and Economics (BME)
  • Semmelweis University (SE)