Data Augmentation with Variational Autoencoders and Manifold Sampling - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Data Augmentation with Variational Autoencoders and Manifold Sampling

Résumé

We propose a new efficient way to sample from a Variational Autoencoder in the challenging low sample size setting. This method reveals particularly well suited to perform data augmentation in such a low data regime and is validated across various standard and real-life data sets. In particular, this scheme allows to greatly improve classification results on the OASIS database where balanced accuracy jumps from 80.7\% for a classifier trained with the raw data to 88.6\% when trained only with the synthetic data generated by our method. Such results were also observed on 3 standard data sets and with other classifiers. A code is available at \url{https://github.com/clementchadebec/Data_Augmentation_with_VAE-DALI}.
Fichier principal
Vignette du fichier
DA_with_VAE_and_manifold_sampling.pdf (4.07 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03180775 , version 1 (25-03-2021)
hal-03180775 , version 2 (21-10-2021)

Identifiants

  • HAL Id : hal-03180775 , version 2

Citer

Clément Chadebec, Stéphanie Allassonnière. Data Augmentation with Variational Autoencoders and Manifold Sampling. DALI 2021 : 1st MICCAI Workshop on Data Augmentation, Labeling, and Imperfections, Oct 2021, Strasbourg, France. ⟨hal-03180775v2⟩
334 Consultations
496 Téléchargements

Partager

Gmail Facebook X LinkedIn More