Disentangled latent representations of images with atomic autoencoders
NEWSON, Alasdair
Laboratoire Traitement et Communication de l'Information [LTCI]
Département Images, Données, Signal [IDS]
Image, Modélisation, Analyse, GEométrie, Synthèse [IMAGES]
Laboratoire Traitement et Communication de l'Information [LTCI]
Département Images, Données, Signal [IDS]
Image, Modélisation, Analyse, GEométrie, Synthèse [IMAGES]
NEWSON, Alasdair
Laboratoire Traitement et Communication de l'Information [LTCI]
Département Images, Données, Signal [IDS]
Image, Modélisation, Analyse, GEométrie, Synthèse [IMAGES]
< Reduce
Laboratoire Traitement et Communication de l'Information [LTCI]
Département Images, Données, Signal [IDS]
Image, Modélisation, Analyse, GEométrie, Synthèse [IMAGES]
Language
en
Communication dans un congrès
This item was published in
Sampling Theory and Applications Conference, 2023-07-10, Sampling Theory and Applications Conference. 2023
English Abstract
We present the atomic autoencoder architecture, which decomposes an image as the sum of elementary parts that are parametrized by simple separate blocks of latent codes. We show that this simple architecture is induced ...Read more >
We present the atomic autoencoder architecture, which decomposes an image as the sum of elementary parts that are parametrized by simple separate blocks of latent codes. We show that this simple architecture is induced by the definition of a general atomic low-dimensional model of the considered data. We also highlight the fact that the atomic autoencoder achieves disentangled low-dimensional representations under minimal hypotheses. Experiments show that their implementation with deep neural networks is successful at learning disentangled representations on two different examples: images constructed with simple parametric curves and images of filtered off-the-grid spikes.Read less <
English Keywords
low dimensional models
autoencoders
disentanglement
ANR Project
Régularisation performante de problèmes inverses en grande dimension pour le traitement de données - ANR-20-CE40-0001
Édition d'imgage avec des réseaux génératifs profonds - ANR-21-CE23-0024
Édition d'imgage avec des réseaux génératifs profonds - ANR-21-CE23-0024
Origin
Hal imported