AVAE: Adversarial Variational Auto Encoder
This work aims to improve the realism of images generated by VAEs for researchers and practitioners in generative modeling, offering an incremental improvement by combining existing methods.
This paper addresses the issue of VAEs generating less realistic images than GANs, attributing it to an underestimation of the natural image manifold dimensionality. The authors propose a new framework, AVAE, that combines VAE and GAN to produce an auto-encoding model capable of generating GAN-quality images while retaining VAE properties.
Among the wide variety of image generative models, two models stand out: Variational Auto Encoders (VAE) and Generative Adversarial Networks (GAN). GANs can produce realistic images, but they suffer from mode collapse and do not provide simple ways to get the latent representation of an image. On the other hand, VAEs do not have these problems, but they often generate images less realistic than GANs. In this article, we explain that this lack of realism is partially due to a common underestimation of the natural image manifold dimensionality. To solve this issue we introduce a new framework that combines VAE and GAN in a novel and complementary way to produce an auto-encoding model that keeps VAEs properties while generating images of GAN-quality. We evaluate our approach both qualitatively and quantitatively on five image datasets.