Parallel Imaging Reconstruction

Parallel imaging reconstruction aims to accelerate magnetic resonance imaging (MRI) acquisition by undersampling k-space data and then reconstructing the full image. Current research heavily utilizes deep learning, particularly generative models (e.g., GANs, score-based models) and implicit neural representations, often incorporating techniques like low-rank tensor decomposition and Hankel matrix transformations to improve reconstruction quality and robustness. These advancements enable faster scans and higher-resolution images, impacting clinical workflows and potentially expanding the accessibility of MRI.