Dose Positron Emission Tomography

Dose-reduction in Positron Emission Tomography (PET) aims to minimize radiation exposure to patients while maintaining diagnostic image quality. Current research heavily utilizes deep learning, employing various architectures like generative adversarial networks (GANs), diffusion models, and transformer networks to reconstruct high-quality standard-dose PET images from low-dose scans. These methods often incorporate multi-modal data (e.g., MRI, CT) or leverage self-supervised learning and prior knowledge to improve reconstruction accuracy and robustness. This field is crucial for advancing patient safety and expanding the clinical applications of PET imaging.