JOINT DISENTANGLEMENT OF LABELS AND THEIR FEATURES WITH VAE - Archive ouverte HAL Access content directly
Conference Papers Year : 2022

JOINT DISENTANGLEMENT OF LABELS AND THEIR FEATURES WITH VAE

Abstract

Most of previous semi-supervised methods that seek to obtain disentangled representations using variational autoencoders divide the latent representation into two components: the non-interpretable part and the disentangled part that explicitly models the factors of interest. With such models, features associated with high-level factors are not explicitly modeled, and they can either be lost, or at best entangled in the other latent variables, thus leading to bad disentanglement properties. To address this problem, we propose a novel conditional dependency structure where both the labels and their features belong to the latent space. We show using the CelebA dataset that the proposed model can learn meaningful representations, and we provide quantitative and qualitative comparisons with other approaches that show the effectiveness of the proposed method.
Fichier principal
Vignette du fichier
ICIP2022.pdf (3.82 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03780425 , version 1 (19-09-2022)

Identifiers

  • HAL Id : hal-03780425 , version 1

Cite

Kaifeng Zou, Sylvain Faisan, Fabrice Heitz, Sebastien Valette. JOINT DISENTANGLEMENT OF LABELS AND THEIR FEATURES WITH VAE. International Conference on Image Processing 2022, Oct 2022, Bordeaux, France. ⟨hal-03780425⟩
66 View
61 Download

Share

Gmail Facebook X LinkedIn More