Fusion U and V: Efficient MRI Sampling via the Fusion of U-net and Conditional Variational Autoencoder
Khristina Pershina, Ching Chun Huang · 2024
Recent years have seen increased interest in Accelerated Magnetic Resonance Imaging (MRI) techniques, valued for their capacity to shorten scan durations while maintaining image fidelity. Our paper introduces a novel approach, Fusion UandV, which employs a Conditional Variational Autoencoder (cVAE) alongside a U-Net architecture to optimize MRI sampling, ensuring accurate reconstruction by conditioning the cVAE model on specific sampling patterns.