Generating depth images of preterm infants in given poses using GANs

Spatio-temporal layers based intra-operative stereo depth estimation network via hierarchical prediction and progressive training

“…Thanks to the contributions of previous scholars in making annotated medical datasets such as Endoabs [45], SERV-CT [46], and the toolkit of Vision Blender [41], which could promote the development of supervised learning in the medical community. A possible improvement is to train our network without the requirement of annotated datasets, for instance, by introducing GAN to train the network in an adversarial way [8,47], which may release the burden of requiring annotated medical datasets.…”