Positron Emission Tomography
Positron Emission Tomography (PET) is a medical imaging technique used to visualize and quantify metabolic processes within the body, primarily for cancer diagnosis and treatment planning. Current research heavily focuses on improving image quality while minimizing radiation exposure, employing deep learning models like U-Nets, transformers, and diffusion models for tasks such as image reconstruction, lesion segmentation, and tracer conversion. These advancements aim to enhance diagnostic accuracy, streamline workflows, and ultimately improve patient care by providing more precise and efficient imaging data.
Papers
STPDnet: Spatial-temporal convolutional primal dual network for dynamic PET image reconstruction
Rui Hu, Jianan Cui, Chengjin Yu, Yunmei Chen, Huafeng Liu
DULDA: Dual-domain Unsupervised Learned Descent Algorithm for PET image reconstruction
Rui Hu, Yunmei Chen, Kyungsang Kim, Marcio Aloisio Bezerra Cavalcanti Rockenbach, Quanzheng Li, Huafeng Liu
Multimodal Deep Learning to Differentiate Tumor Recurrence from Treatment Effect in Human Glioblastoma
Tonmoy Hossain, Zoraiz Qureshi, Nivetha Jayakumar, Thomas Eluvathingal Muttikkal, Sohil Patel, David Schiff, Miaomiao Zhang, Bijoy Kundu
Self-Supervised Pre-Training for Deep Image Prior-Based Robust PET Image Denoising
Yuya Onishi, Fumio Hashimoto, Kibo Ote, Keisuke Matsubara, Masanobu Ibaraki