Fig. 6: Overview of presented methods for the generation of nuclei images based on simulated cell labels.
From: Improving 3D deep learning segmentation with biophysically motivated cell synthesis
SimOptiGAN+ (A) uses the simulated cell borders to improve the nuclei placement of a previously presented simulation-based pipeline, while Mem2NucGAN-P (B) and Mem2NucGAN-U (C) rely on a GAN-based transformation of simulated cell borders.
