Resolution improvement on PET images by using wobble motion and super-resolution processing = Wobble 모션과 초고해상도 처리에 기반한 PET 영상의 해상도 향상 기법

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 401
  • Download : 0
DC FieldValueLanguage
dc.contributor.advisorRa, Jong-Beom-
dc.contributor.authorJeong, Kye-Young-
dc.description학위논문(박사) - 한국과학기술원 : 전기및전자공학전공, 2012.8, [ ix, 93 p. ]-
dc.description.abstractPositron emission tomography (PET) images have been increasingly used in oncology, cardiology, and neurology, because they can provide quantitative and qualitative metabolic data for living organisms. However, PET images suffer from low spatial resolution. The resolution improvement can be partially achieved by increasing the number of samples through a scanner motion such as wobble and dichotomic motions. To further increase the spatial resolution of PET images, super-resolution (SR) was attempted in the sinogram space or the image space. A SR algorithm can produce a high-resolution (HR) image (or sinogram) from multiple sinograms or multiple low-resolution (LR) images. However, SR in PET is still a challenging problem. This dissertation deals with a sinogram-based SR algorithm for two kinds of PET scanners: whole-body scanner and small animal scanner. In the whole-body PET scanner with a large system diameter, measured sinograms can be assumed to have a space invariant blur kernel. Based on this assumption, we propose a novel and efficient sinogram-based SR algorithm. In the algorithm, a blur kernel is estimated through a Monte Carlo simulation. A HR sinogram is then determined from the sub-pixel shifted sinograms obtained through a discrete wobble motion. In this one dimensional SR process, we adopt a maximum a posteriori expectation maximization (MAP-EM) method based on the estimated blur kernel. We finally reconstruct a HR image from the HR sinogram through the conventional reconstruction algorithm. Experiments were performed by using a resolution phantom and a line source in a commercial whole-body PET scanner, ECAT HR+. The results show that the proposed algorithm provides noticeable improvement on the spatial resolution. The full width at half-maximum value is reduced from 4.4 mm to 2.2 mm. In a small animal PET scanner with a small system diameter, however, a space invariant blur kernel is not appropriate for providing a HR sinogram, because the spat...eng
dc.subjectPositron emission tomography (PET)-
dc.titleResolution improvement on PET images by using wobble motion and super-resolution processing = Wobble 모션과 초고해상도 처리에 기반한 PET 영상의 해상도 향상 기법-
dc.identifier.CNRN511889/325007 -
dc.description.department한국과학기술원 : 전기및전자공학전공, -
dc.contributor.localauthorRa, Jong-Beom-
Appears in Collection
Files in This Item
There are no files associated with this item.


  • mendeley


rss_1.0 rss_2.0 atom_1.0