Generalized Latent Variable Recovery for Generative Adversarial Networks
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Generalized Latent Variable Recovery for Generative Adversarial Networks

Authors: Nicholas Egan, Jeffrey Zhang, Kevin Shen

The Generator of a Generative Adversarial Network (GAN) is trained to transform latent vectors drawn from a prior distribution into realistic looking photos. These latent vectors have been shown to encode information about the content of their corresponding images. Projecting input images onto the latent space of a GAN is non-trivial, but previous work has successfully performed this task for latent spaces with a uniform prior. We extend these techniques to latent spaces with a Gaussian prior, and demonstrate our technique's effectiveness.

See our paper for details.

Our GAN was trained on food photos from the Yelp dataset and uses the DCGAN architecture in PyTorch.

Generator Samples

Standard DCGAN

Pixel Shuffle DCGAN

Soft Label DCGAN

Latent Vector Interpolation (SLERP)