How good is my GAN?

Konstantin Shmelkov 1 Cordelia Schmid 1 Karteek Alahari 1
1 Thoth - Apprentissage de modèles à partir de données massives
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann
Abstract : Generative adversarial networks (GANs) are one of the most popular methods for generating images today. While impressive results have been validated by visual inspection, a number of quantitative criteria have emerged only recently. We argue here that the existing ones are insufficient and need to be in adequation with the task at hand. In this paper we introduce two measures based on image classification---GAN-train and GAN-test, which approximate the recall (diversity) and precision (quality of the image) of GANs respectively. We evaluate a number of recent GAN approaches based on these two measures and demonstrate a clear difference in performance. Furthermore, we observe that the increasing difficulty of the dataset, from CIFAR10 over CIFAR100 to ImageNet, shows an inverse correlation with the quality of the GANs, as clearly evident from our measures.
Type de document :
Communication dans un congrès
Vittorio Ferrari; Martial Hebert; Cristian Sminchisescu; Yair Weiss. ECCV 2018 - European Conference on Computer Vision, Sep 2018, Munich, Germany. pp.1-20, 2018, Lecture Notes in Computer Science
Liste complète des métadonnées

https://hal.inria.fr/hal-01850447
Contributeur : Karteek Alahari <>
Soumis le : vendredi 27 juillet 2018 - 13:15:33
Dernière modification le : jeudi 13 septembre 2018 - 09:54:18
Document(s) archivé(s) le : dimanche 28 octobre 2018 - 13:24:16

Fichier

ganeval.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01850447, version 1
  • ARXIV : 1807.09499

Collections

Citation

Konstantin Shmelkov, Cordelia Schmid, Karteek Alahari. How good is my GAN?. Vittorio Ferrari; Martial Hebert; Cristian Sminchisescu; Yair Weiss. ECCV 2018 - European Conference on Computer Vision, Sep 2018, Munich, Germany. pp.1-20, 2018, Lecture Notes in Computer Science. 〈hal-01850447〉

Partager

Métriques

Consultations de la notice

1118

Téléchargements de fichiers

359