Meeting Banner
Abstract #3960

Exploring RF pulse design with deep reinforcement learning

Xiaodong Ma1, Kamil U─čurbil1, and Xiaoping Wu1
1Center for Magnetic Resonance Research, Radiology, Medical School, University of Minnesota, Minneapolis, MN, United States

In this study, we expand the application of a deep reinforcement learning (DRL) pulse design framework to designing four basic types of RF pulses and more complicated multi-band RF pulses. Our results showed that the DRL framework can be used to effectively design all types of RF pulses, improving slice profiles with reduced ripple levels in comparison to the conventional SLR algorithm.

This abstract and the presentation materials are available to members only; a login is required.

Join Here