Recognizing Art Style Automatically in Painting with Deep Learning

Adrian Lecoutre, Benjamin Negrevergne, Florian Yger
Proceedings of the Ninth Asian Conference on Machine Learning, PMLR 77:327-342, 2017.

Abstract

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.

Cite this Paper

BibTeX
@InProceedings{pmlr-v77-lecoutre17a, title = {Recognizing Art Style Automatically in Painting with Deep Learning}, author = {Lecoutre, Adrian and Negrevergne, Benjamin and Yger, Florian}, booktitle = {Proceedings of the Ninth Asian Conference on Machine Learning}, pages = {327--342}, year = {2017}, editor = {Zhang, Min-Ling and Noh, Yung-Kyun}, volume = {77}, series = {Proceedings of Machine Learning Research}, address = {Yonsei University, Seoul, Republic of Korea}, month = {15--17 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v77/lecoutre17a/lecoutre17a.pdf}, url = {https://proceedings.mlr.press/v77/lecoutre17a.html}, abstract = {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.} }
Endnote
%0 Conference Paper %T Recognizing Art Style Automatically in Painting with Deep Learning %A Adrian Lecoutre %A Benjamin Negrevergne %A Florian Yger %B Proceedings of the Ninth Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Min-Ling Zhang %E Yung-Kyun Noh %F pmlr-v77-lecoutre17a %I PMLR %P 327--342 %U https://proceedings.mlr.press/v77/lecoutre17a.html %V 77 %X 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.
APA
Lecoutre, A., Negrevergne, B. & Yger, F.. (2017). Recognizing Art Style Automatically in Painting with Deep Learning. Proceedings of the Ninth Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 77:327-342 Available from https://proceedings.mlr.press/v77/lecoutre17a.html.

Related Material