Please use this identifier to cite or link to this item: https://knowledgecommons.lakeheadu.ca/handle/2453/5025
Full metadata record
DC FieldValueLanguage
dc.contributor.advisorFadlullah, Zubair Md.-
dc.contributor.advisorReznik, Alla-
dc.contributor.authorKomorov, Borys-
dc.date.accessioned2022-09-28T19:37:38Z-
dc.date.available2022-09-28T19:37:38Z-
dc.date.created2022-
dc.date.issued2022-
dc.identifier.urihttps://knowledgecommons.lakeheadu.ca/handle/2453/5025-
dc.description.abstractFunctional medical imaging is unique in its ability to visualize molecular interactions and pathways in the body. Organ-targeted Positron Emission Tomography (PET) is a functional imaging technique that has emerged to meet the demands of precision medicine and has shown advantages in terms of sensitivity and image quality compared to whole-body (WB) PET. A common application for organ-targeted PET is oncology, particular breast cancer imaging. In this work we present the application of Graphics Processing Unit (GPU) to significantly accelerate reconstruction of clinical breast images acquired with an organ-targeted PET camera and reconstructed using the Maximum Likelihood Estimation Maximization (MLEM) algorithm. The PET camera is configured with two planar detector heads with a sensing area of 232mm×174mm. Acquired raw image data are converted into list mode format and reconstructed by a GPU-based 3D MLEM algorithm that was developed specifically for the limited-angle capabilities of the planar PET geometry. The algorithm applies corrections including attenuation and scatter to provide clinical grade image quality. We demonstrate that a transition from originally developed Central Processing Unit (CPU) to newly developed GPU-based algorithm improves single iteration speed by more than 400 times, while preserving image quality. The latter has been assessed on clinical data and through phantom tests performed according to the National Electrical Manufacturers Association (NEMA) NU-4 standards. The gain in reconstruction speed is expected to result in improved patient throughput capabilities of the clinical organ-targeted PET. Indeed, GPU-based image reconstruction reduces time needed for a typical breast image reconstruction to less than 5 minutes thus making it shorter than the image acquisition time. This is of particular importance in improving patient throughput and clinical adaptation of the PET system.en_US
dc.language.isoen_USen_US
dc.subjectPositron Emission Tomography (PET)en_US
dc.subjectGraphics Processing Unit (GPU)en_US
dc.subjectBreast imagingen_US
dc.title3D GPU-based image reconstruction algorithm for the application in a clinical organ-targeted PET cameraen_US
dc.typeThesisen_US
etd.degree.nameMaster of Scienceen_US
etd.degree.levelMasteren_US
etd.degree.disciplineComputer Scienceen_US
etd.degree.grantorLakehead Universityen_US
Appears in Collections:Electronic Theses and Dissertations from 2009

Files in This Item:
File Description SizeFormat 
KomarovB2022m-1a.pdf1.46 MBAdobe PDFThumbnail
View/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.