Author identification for digitized paintings collections

This paper presents an automatic system for the painter recognition from digital representations of paintings. The proposed solution comes as part of the recent extensive effort of developing image processing solutions that facilitate a better understanding of art. Each painting is described with low-level features motivated by art theory (3D RGB Histograms and Gabor Energy Features). The paper presents the possible use of eight classifiers, the best performance being obtained using a Multi Class Classifier. The system's performance is evaluated on a database containing 1800 paintings belonging to 15 different painters, proving to outperform the reported state of the art.