Prediction of human eye fixations using symmetry

Humans are very sensitive to symmetry in visual patterns. Reaction time experiments show that symmetry is detected and recognized very rapidly. This suggests that symmetry is a highly salient feature. Existing computational models of saliency, however, have mainly focused on contrast as a measure of saliency. In this paper, we discuss local symmetry as a measure of saliency. We propose a number of symmetry models and perform an eye-tracking study with human participants viewing photographic images to test the models. The performance of our symmetry models is compared with the contrast-saliency model of Itti, Koch and Niebur (1998). The results show that the symmetry models better match the human data than the contrast model, which indicates that symmetry can be regarded as a salient feature