Use of Digital Technology for the Attribution of Paintings

Abstract

Modern attribution of art works depends on the expert experience. This implies some subjectivity of his assessments. Recently for making a more informed decision digital image processing techniques have been used. Therefore, the study of the methods for obtaining additional information about the painting, which cannot be obtained using existing methods, is an important task. The joint use of several features extraction methods will allow increase the effectiveness of its analysis. The application developed by the authors based on using of several methods (HOG, DCT, frequency analysis of color, color moments, color histograms, LBP, Gabor filter, GLCM, Hough transform) and allows making an assumption about belonging artwork to a particular art school, style or direction. The results obtained are confirmed by historical data about his teacher and where the artist studied, and who of the masters had the greatest influence on him.