Classification of art paintings by genre

This paper offers an approach to automatic art genre classification of paintings. Development of machine learning algorithms and increase of overall computing power improved speed and efficiency of feature extraction from digital images and with it opened a whole new set of possibilities in classification of visual data such as paintings and other visual art. Automatic classification is useful in large database processing (e.g. museums) and could be used as a commercial application on mobile platforms. Six genres are classified in the paper: realism, impressionism, cubism, fauvism, pointilism and naïve art. Some of the genres have now been tested for the first time. Used features are described as well as a measure of their usefulness. Rate of success for different classifiers is given. Accomplished results are similar to related work results.