Convolutional Symmetric AutoEncoders: enhancing latent stability via differential geometry

Open in new window