Generating depth images of preterm infants in given poses using GANs

Giuseppe Pio Cannata, Lucia Migliorelli, Adriano Mancini, Emanuele Frontoni, Rocco Pietrini, Sara Moccia

Abstract: Highlights•Datasets lack limits the spread of deep-learning algorithms to monitor preterm infants.•We propose a GAN framework to generate realistic depth images of preterm infants.•Validation is performed on the Moving INfants In RGB-D dataset with 12 infants.•Quantitative and qualitative evaluation of the framework shows promising results.