Learning Visual Balance from Large-scale Datasets of Aesthetically Highly Rated Images

The concept of visual balance is innate for humans, and influences how we perceive visual aesthetics and cognize harmony. Although visual balance is a vital principle of design and taught in schools of designs, it is barely quantified. On the other hand, with emergence of automantic/semi-automatic visual designs for self-publishing, learning visual balance and computationally modeling it, may escalate aesthetics of such designs. In this paper, we present how questing for understanding visual balance inspired us to revisit one of the well-known theories in visual arts, the so called theory of "visual rightness", elucidated by Arnheim. We define Arnheim's hypothesis as a design mining problem with the goal of learning visual balance from work of professionals. We collected a dataset of 120K images that are aesthetically highly rated, from a professional photography website. We then computed factors that contribute to visual balance based on the notion of visual saliency. We fitted a mixture of Gaussians to the saliency maps of the images, and obtained the hotspots of the images. Our inferred Gaussians align with Arnheim's hotspots, and confirm his theory. Moreover, the results support the viability of the center of mass, symmetry, as well as the Rule of Thirds in our dataset.