generative models capable of capturing nuanced clinical features in medical
images hold great promise for facilitating clinical data sharing, enhancing
rare disease datasets, and efficiently synthesizing annotated medical images at
scale. Despite their potential, assessing the quality