Prospective Dynamic 3D MRI Reconstruction via Latent-Space Motion Tracking from Single Measurement
Researchers have developed a new framework called PDMR for prospective dynamic 3D MRI reconstruction. This method utilizes latent-space motion tracking to efficiently estimate motion fields from sparse MRI data, which is crucial for applications like MRI-guided radiotherapy. PDMR learns a low-dimensional manifold of motion fields offline, allowing for rapid online adaptation and high-fidelity reconstruction, outperforming existing retrospective and online techniques in experiments. AI