/vae_latent_space_viewer

A simple viewer for exploring the latent space of Variational Autoencoder (VAE) models.

Primary LanguagePython

VAE Latent Space Viewer

A simple viewer for exploring the latent space of Variational Autoencoder models. This is just a small demo that comes along with the VAE Pancreas Segmentation repository.

Screenshot of the viewer

A model is provided, trained on pancreas segmentations from the TCIA Pancreas-CT dataset. A random sample from the validation dataset was chosen to be encoded to a latent spacer vector z as the center for exploration.