This paper tests the effectiveness of GAN-created ultrasound images to augment small training sets, evaluated by performance of a downstream classifier of pneumonia features.

Published in the Sashimi workshop, MICCAI, 2023.

A local copy is here.