Recognizing Art Style Automatically with deep learning - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Proceedings of Machine Learning Research Année : 2017

Recognizing Art Style Automatically with deep learning

Résumé

The artistic style (or artistic movement) of a painting is a rich descriptor that captures both visual and historical information about the painting. Correctly identifying the artistic style of a paintings is crucial for indexing large artistic databases. In this paper, we investigate the use of deep residual neural to solve the problem of detecting the artistic style of a painting and outperform existing approaches to reach an accuracy of 62% on the Wikipaintings dataset (for 25 different style). To achieve this result, the network is first pre-trained on ImageNet, and deeply retrained for artistic style. We empirically evaluate that to achieve the best performance, one need to retrain about 20 layers. This suggests that the two tasks are as similar as expected, and explain the previous success of hand crafted features. We also demonstrate that the style detected on the Wikipaintings dataset are consistent with styles detected on an independent dataset and describe a number of experiments we conducted to validate this approach both qualitatively and quantitatively.
Fichier principal
Vignette du fichier
lecoutre17a.pdf (1.02 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02004781 , version 1 (02-02-2019)

Identifiants

  • HAL Id : hal-02004781 , version 1

Citer

Adrian Lecoutre, Benjamin Negrevergne, Florian Yger. Recognizing Art Style Automatically with deep learning. Proceedings of Machine Learning Research, 2017, 77, pp.327 - 342. ⟨hal-02004781⟩
176 Consultations
1959 Téléchargements

Partager

Gmail Facebook X LinkedIn More