A new symmetry measure based on Gabor filters

Symmetry is one of the important factors affecting aesthetic quality of an artwork. In paintings symmetry is understood as balanced arrangement of objects rather than a perfect mirrored symmetry. Computational aesthetic which was first introduced by Birkhof involves assigning values to each component of aesthetic measure. Processing power of today's computers as well as image processing tools has paved the way for automatic and objective evaluation of paintings. In this research, Gabor filters are used to devise a new symmetry measure appropriate for computational evaluation of paintings. This symmetry measure is computed for various images from synthetic and natural images to images of miniature paintings. Comparing this measure and two previously used measures against the subjective measure of symmetry confirms the effectiveness of our proposed measure. The results show a good correlation between the perceived symmetry of a painting and its computed symmetry measure.