Positron emission tomography (PET) is an indispensable medical imaging technique which can reveal the physiology activities by injecting special PET tracers. In recent years, deep learning has been widely used in PET reconstruction tasks. In this work, a multiscale direct reconstruction framework based on cGANs has been presented. First, the generator is built as two scales which will combine both global and detail information in training phase. Second, to match the generator, we tried multiscale discriminators. Each discriminator shares the same structure, but operates on different scales. Both coarse field and fine field were scored by the discriminators, so this score considered not only the global structure but also fine details. In order to verify the feasibility of the proposed framework, both simulation datasets and real SD rat dataset were utilized in our experiments. We also compared our method with U-Net and Deep- PET network, the results showed that we got more accurate and real radioactivity images on all datasets.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.