Guessing Art Styles using Qualitative Colour Descriptors, SVMs and Logics
Falomir Zoe, Lledó Museros, Ismael Sanz, Luis González Abril · Frontiers in artificial intelligence and applications · 2015
An approach for style painting categorization based on their Qualitative Colour Descriptors (QCD) and their similarity (SimQCD) is presented in this paper. For learning the colour features of paintings from Baroque, Impressionism and Post-Impressionism styles, two classifiers are built based on k-Nearest Neighbour (k-NN) and support vector machine (SVM) methods. The results showed accuracies of categorization higher than 75%. Colour logics in art styles are identified from the literature in order to improve and explain the categorization obtained.