AVAE: Adversarial variational auto encoder - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

AVAE: Adversarial variational auto encoder

Résumé

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.
Fichier principal
Vignette du fichier
2012.11551.pdf (8.48 Mo) Télécharger le fichier

Dates et versions

hal-04314595 , version 1 (29-11-2023)

Identifiants

Citer

Antoine Plumerault, Hervé Le Borgne, Céline Hudelot. AVAE: Adversarial variational auto encoder. ICPR 2020 - 25th International conference on pattern recognition, Jan 2021, Milan, Italy. pp.8687-8694, ⟨10.1109/ICPR48806.2021.9412727⟩. ⟨hal-04314595⟩
14 Consultations
19 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More