The DELAUNAY dataset for psychophysical and machine learning experiments

Visualization of DELAUNAY dataset.

The DELAUNAY (Dataset for Experiments on Learning with Abstract and Non-figurative art for Neural Networks and Artificial intelligence) dataset is named after artists Sonia and Robert Delaunay.

Contents

The dataset comprises 11,503 images from 53 categories, i.e. artists (mean number of images per artist: 217.04; standard deviation: 58.55), along with the associated URLs.

Citing this data set

Please cite the following paper:

C. Gontier, J. Jordan, and M. A. Petrovici (2022). DELAUNAY: a dataset of abstract art for psychophysical and machine learning research. arXiv preprint.

Images available at https://physiologie.unibe.ch/preprints.aspx