Newspaper Layout Aesthetics Judged by Artificial Neural Networks

Automating the process of newspaper page composition is a complicated task especially because it involves an automatic evaluation of layout aesthetics. Dealing with aesthetics is a major problem since it does not seem possible to describe the concept of aesthetics by a few simple rules. It is investigated whether artificial neural networks are suited for the task of evaluating layout aesthetics. Based on interviews with typographers, two criteria of aesthetics: placement of articles with respect to their size and priority and distribution of weight, have been chosen for further investigations. It was not possible to obtain satisfactory performance with an artificial neural network judging weight distribution. This failure was later shown to be caused by tacit knowledge used by typographers in their evaluation of layouts. The experiences drawn from these experiments indicate that artificial neural networks are not suited for the task of evaluation of layout aesthetics. But the knowledge acquired through the project form the basis of a framework for use in further work towards an automation of newspaper page composition. ∗This work was granted by CCI Europe and by the ESPRIT Long Term Research Programme of the EU under project number 20244 (ALCOM-IT).