FetalBrainAwareNet: Bridging GANs with anatomical insight for fetal ultrasound brain plane synthesis

Angelo Lasala, Maria Chiara Fiorentino, Andrea Bandini, Sara Moccia

Abstract: Highlights•Novel conditional GAN framework developed for fetal ultrasound image synthesis.•This framework uses anatomical-aware regularization with class activation maps as prior condition.•First framework to synthesize various categories of fetal ultrasound images together.•Framework boosts classifiers with varied image synthesis used as data augmentation.