Cross Domain Image Generation through Latent Space Exploration with Adversarial Loss
This addresses the challenge of flexible conditional generation in deep generative models for image synthesis applications, though it appears incremental as it builds on existing VAE and domain transfer techniques.
The paper tackles the problem of expensive adaptation of conditional generative models to new controls by proposing a conditioned latent domain transfer framework that enables unconditionally trained VAEs to generate images with conditionals from another domain's latent representation, demonstrating effectiveness on widely used image datasets.
Conditional domain generation is a good way to interactively control sample generation process of deep generative models. However, once a conditional generative model has been created, it is often expensive to allow it to adapt to new conditional controls, especially the network structure is relatively deep. We propose a conditioned latent domain transfer framework across latent spaces of unconditional variational autoencoders(VAE). With this framework, we can allow unconditionally trained VAEs to generate images in its domain with conditionals provided by a latent representation of another domain. This framework does not assume commonalities between two domains. We demonstrate effectiveness and robustness of our model under widely used image datasets.